federal university of ouro preto - repositorio.ufop.br‡Ão... · ii federal university of ouro...
TRANSCRIPT
I
FEDERAL UNIVERSITY OF OURO PRETO
GRADUATE PROGRAMME ON MATERIALS ENGINEERING
Master’s Thesis
“Effect of Ti and Nb Additions on the Formation of Craters for IF Steel
Galvanneal Coatings”
Author: Breno Storch
Supervisors: Dr Fernando Gabriel S. Araújo
Dr Joseph McDermid
November, 2009
II
FEDERAL UNIVERSITY OF OURO PRETO
GRADUATE PROGRAMME ON MATERIALS ENGINEERING
Breno Storch
EFFECT OF Ti AND Nb ADDITIONS ON THE FORMATION OF CRATERS FOR
IF STEEL GALVANNEAL COATINGS
Master’s Thesis presented to the Materials Engineering
Graduate Programme at the Federal University of Ouro
Preto, Brazil
Research Area: Coating Developments
Supervisors: Dr Fernando Gabriel S. Araújo
Dr Joseph McDermid
Ouro Preto, Brazil
Federal University of Ouro Preto
2009
3
MASTER’S THESIS TABLE OF CONTENTS
List of Figures ....................................................................................................................... 5
List of Tables......................................................................................................................... 7
Abstract................................................................................................................................. 8
1. INTRODUCTION ........................................................................................................... 9
2. LITERATURE BACKGROUND.................................................................................... 10
2.1. INFLUENCE OF ALLOYING ELEMENTS ON IF STEELS....................................... 10
2.2. C, N, S AND P STABILIZATION IN IF STEELS ....................................................... 11
2.3. MICROSTRUCTURE OF IF STEELS....................................................................... 13
2.4. CONTINUOUS GALVANIZING AND GALVANNEALING......................................... 14
2.4.1. FORMATION OF GALVANNEALED COATINGS.............................................. 16
2.4.2. GROWTH OF Fe-Zn INTERMETALLIC COMPOUNDS .................................... 30
2.4.3. CRATER FORMATION...................................................................................... 49
2.4.4. EFFECT OF IF STEEL ALLOYING ELEMENTS ON GALVANNEALED
COATING STRUCTURE ............................................................................................. 53
3. EXPERIMENTAL PROCEDURE .................................................................................... 62
3.1. MATERIALS ............................................................................................................. 62
3.2. TESTS PERFORMED.............................................................................................. 64
3.3. METALLOGRAPHIC CHARACTERIZATION OF AS-RECEIVED SAMPLES. ......... 65
3.4. IF STEEL GA SAMPLES WITH STREAKS.............................................................. 65
3.5. HOT-DIP PROCESS SIMULATOR .......................................................................... 66
3.5.1. DEFINITION OF HEAT TREATMENT CYCLES FOR Ti- AND DUAL-
STABILIZED IF STEEL GRADES................................................................................ 67
3.5.2. INTERRUPTED CYCLES FOR IF STEEL BARE SAMPLES ............................ 68
3.5.3. Ti-STABILIZED IF STEEL GRADE.................................................................... 69
3.5.4. DUAL-STABILIZED IF STEEL GRADE ............................................................. 69
4
3.5.5. POST-PROCESSING AND STORAGE OF EXPERIMENTAL COUPONS ....... 70
3.6. SAMPLE ANALYSIS TECHNIQUES........................................................................ 70
4. RESULTS AND DISCUSSION........................................................................................ 74
4.1. MICROSTRUCTURE CHARACTERIZATION.............................................................. 74
4.2. CHEMICAL PROFILES ON Ti- AND DUAL-STABILIZED IF STEEL BARE AND
COATED SAMPLES ....................................................................................................... 75
4.3. CRATER ANALYSIS ON TOP OF GA COATINGS.................................................. 88
4.3.1. CRATER ANALYSIS ON LIGHT AND DARK STREAKED AREAS FOR AS-
RECEIVED GA COATINGS......................................................................................... 88
4.3.2. SAMPLES RUN UNDER SPECIAL COOLING CONDITIONS AT THE HOT-DIP
PROCESS SIMULATOR ............................................................................................. 90
4.4. STEEL SUBSTRATE REACTIVITY ALONG FERRITE GRAIN BOUNDARIES ON Ti-
AND DUAL-STABILIZED IF STEEL SUBSTRATES: ...................................................... 92
4.4.1. REGIONS OF HIGH REACTIVITY AND FORMATION OF OUTBURSTS ........ 93
4.4.2. REGIONS OF HIGH REACTIVITY AND FORMATION OF CRATERS ON LIGHT
AND DARK STREAKED AREAS................................................................................. 93
4.5. THE EFFECT OF THE DIFFUSION OF IF STEEL ALLOYING ELEMENTS
TOWARDS THE INTERFACE BETWEEN IF STEEL SUBSTRATES AND GA
COATINGS ON CRATER FORMATION ......................................................................... 94
4.5.1. Ti DEPLETION AT THE STEEL SUBSTRATE-GA COATING INTERFACE; .... 94
4.5.2. Nb ENRICHMENT AT THE STEEL SUBSTRATE-GA COATING INTERFACE 95
4.5.3. EFFECT OF Ti AND Nb SURFACE SEGREGATION ON CRATER FORMATION
..................................................................................................................................... 98
5. CONCLUSIONS............................................................................................................ 100
6. ACKNOWLEDGEMENTS...............................................................................................101
7. BIBLIOGRAPHY ........................................................................................................... 102
5
List of Figures Figure 1. Flowchart of the steps carried out by this research project. ................................. 64
Figure 2. Thermal profiles set up for both IF steel grades at the HDP simulator. The
markers show sampling conditions. .................................................................................... 69
Figure 3. Thermal sequence for sampling with interrupted GA thermal cycles for both IF
steel grades at the HDP simulator. ..................................................................................... 73
Figure 4. Micrograph featuring typical microstructures found for the as-received (as-cold-
rolled) samples for Ti-stabilized IF steel substrate (optical microscope, magnification: 100X)
............................................................................................................................................ 74
Figure 5. Micrograph featuring typical microstructure found for the as-received (as-cold-
rolled) sample for dual-stabilized IF steel substrate (optical microscope, magnification:
100X). ................................................................................................................................. 75
Figure 6. SEM images at 2100x and 2020x featuring the GA coating cross section for the
Ti-stabilized IF steel grade (left-hand side) and the dual-stabilized IF steel grade (right-hand
side). ................................................................................................................................... 76
Figure 7. Spots chosen along the Ti-stabilized IF steel thermal profile. Region within the
solid red square feature the prior condition to sample dipping into the Zn bath. ................. 77
Figure 8. Spots chosen along the dual-stabilized IF steel thermal profile. Region within the
solid red square feature the prior condition to sample dipping into the Zn bath. ................. 77
Figure 9. Ti outermost surface concentration along the spots where bare steel samples
were taken out of Ti-stabilized IF steel thermal profiles. (S1 to 12 => temperature vs. time
condition) ............................................................................................................................ 78
Figure 10. Nb outermost surface concentration along the spots where bare steel samples
were taken out of Ti-stabilized IF steel thermal profiles. (S1 to 12 => temperature vs. time
condition) ............................................................................................................................ 79
Figure 11. Ti outermost surface concentration along the spots where bare steel samples
were taken out of dual-stabilized IF steel thermal profiles. (S1 to 12 => temperature vs.
time condition)..................................................................................................................... 80
Figure 12. Nb outermost surface concentration along the spots where bare steel samples
were taken out of dual-stabilized IF steel thermal profiles. (S1 to 12 => temperature vs.
time condition)..................................................................................................................... 81
6
Figure 13. Ti surface concentration at specific depths results along the thermal cycle set up
for the Ti-stabilized IF steel grade (Ti bulk concentration is 0.062%).................................. 82
Figure 14. Nb surface concentration results at 0.02 and 0.10µm depth along the thermal
cycle set up for the dual-stabilized IF steel grade (Nb bulk concentration is 0.029%)......... 83
Figure 15. XPS surface analysis results for the dual-stabilized IF steel grade, featuring a Nb
3d3/2 peak whose binding energy is 207.4 eV), which corresponds to Nb2O5. ................. 84
Figure 16. GD-OES Zn, Fe and Al atomic concentration across GA coating for a Ti-
stabilized IF steel substrate................................................................................................. 85
Figure 17. GD-OES Zn, Fe and Al atomic concentration across GA coating for a dual-
stabilized IF steel substrate................................................................................................. 86
Figure 18. Ti and Nb concentrations along a 30 µm-thick GA coating on top of a Ti-
stabilized IF steel substrate................................................................................................. 87
Figure 19. Ti and Nb concentrations along a 12 µm-thick GA coating on top of a dual-
stabilized IF steel substrate................................................................................................. 88
Figure 20. Stereoscope micrographs featuring light and dark streaked areas on top of GA
coatings for Ti-stabilized IF steels. Magnification: 6.5 x. ..................................................... 89
Figure 21. SEM micrographs featuring light (right-hand side) and dark (light-hand side)
streaked areas on top of GA coatings for Ti-stabilized IF steels. Magnification: 573 x. ...... 89
Figure 22. SEM images at 2300x featuring typical microstructures for a dark streaked area
(left-hand side), whose EDS chemistry taken from the bottom matches a Γ phase, and light
streaked area (right-hand side), whose chemistry is typical of a δ phase............................ 90
Figure 23. Pictures feature the overall outlook of Ti- (left-hand side) and dual-stabilized
(right-hand side) IF steel panels that run interrupted GA thermal cycles under faster cooling
rates.................................................................................................................................... 91
Figure 24. SEM images and EDS chemistries for craters on bottom of light streaked areas
at GA coatings on top of Ti-stabilized IF steel panels after running interrupted GA thermal
cycles under faster cooling rates......................................................................................... 91
Figure 25. SEM images and EDS chemistries for craters on bottom of dark streaked areas
at GA coatings on top of Ti-stabilized IF steel panels after running interrupted GA thermal
cycles under faster cooling rates......................................................................................... 92
7
List of Tables Table 1. Features of Fe-Zn intermetallic compounds.......................................................... 30
Table 2. Usual chemistry range of Fe-Zn intermetallic compounds. ................................... 31
Table 3. IF Steel substrates and conditions tested. ............................................................ 62
Table 4. Chemistries for the Ti-stabilized IF steel grade and the dual-stabilized IF steel
grade................................................................................................................................... 63
Table 5. Tests performed on the IF steels. ......................................................................... 64
Table 6. HDG operating parameters carried out for setting up thermal cycles run on both IF
steel grades during HDG simulations.................................................................................. 68
Table 7. Ferrite average grain sizes of the IF steels studied before and after galvannealing.
............................................................................................................................................ 75
8
Abstract
Interstitial Free steels feature alloying elements present in an ultra low C matrix that
account for C stabilization and consequently a high number of C free interstitials in a ferrite
matrix that give these steels the ideal texture, low yield point, favorable plastic strain ratios,
high elongation and n-value necessary for an ideal performance on forming press, drawing
and stamping operations especially for automotive body part applications.
This Master’s thesis indicates that Ti-stabilized IF steel substrates are more reactive than
dual-stabilized IF steel substrates probably due to the presence of islands of Nb oxides at
the interface between the steel substrate and the GA coating. These Nb oxides may block
the Fe-Zn interdiffusion along the ferrite grain boundaries. Hence, higher reactivity rates for
Ti-stabilized IF steel grades mean that this substrate is more prone than a dual-stabilized IF
steel to the formation of outbursts and consequently the formation of craters, which are
morphologically characterized as clusters of outbursts, whose formation mechanisms are
based on capillarity effects accounting for amounts of liquid Zn being drained away from
these spots. Besides, this work has also explored the typical chemistry and morphology
features of craters on top of dark and light streaked areas on GA coatings, showing that
there are δ crystals on the bottom of craters on light streaked areas, whereas Γ phase was
found on the bottom of craters on dark streaked areas. Also, it turns out that craters on dark
streaked regions are deeper than those on light streaked areas. On top of that, the crater
coverage on dark streaked regions is larger than on light streaked areas.
9
1. INTRODUCTION
Interstitial Free or IF steels have been the cornerstone of the steel industry for forming,
drawing and stamping applications for a number of years due to their low yield strength,
favorable plastic strain ratios and high uniform elongation.
The industry uses Ti-, Nb- and Ti-Nb-stabilized IF grades for a variety of applications.
These grades are easy to galvanize as the bulk of the substrate is ferrite and preparation of
the substrate surfaces is such that their wettability by liquid Zn-Al alloys becomes trivial.
However, it has been noticed that Ti-containing grades do not behave as well as Nb-
bearing grades when manufacturing galvannealed coatings. The former usually exhibit
features of decreased coating integrity (i.e. decreased coating adherence) and a surface
defect whose appearance is crater-like, which can be characterized as a discontinuity on
the surface of galvanneal coatings and results in poor coating appearance.
These problems are not commonly seen on the surface of Nb-stabilized IF steel coatings.
However, Nb-bearing grades do not generally have the same desirable mechanical
properties (particularly with respect to formability) found in Ti containing grades. The former
are generally prone to have higher alloying costs requiring higher power to recrystallize
anneal and galvanneal. Therefore, Ti-Nb-stabilized IF steel grades were developed as a
compromise between these two behaviors.
It is worth pointing out that the differences between these two grades in the galvannealing
process are remarkable considering the rather low level of alloying elements in these steels
and the fact that the Ti and Nb additions are ‘tied up’ as carbides and nitrides to a large
extent.
It is also worth mentioning that modeling work on the chemical development (i.e. Fe
content) of galvanneal coatings on IF steel substrates has been performed by some
researchers. Some authors have also examined the influence of microstructural and
process variables on the structural formation of metallic coatings on IF steels. However, it is
10
believed that the fundamental origin of the differences between the process behavior of
these steel substrates lies in a reasonable comprehension of the effect of substrate
reactivity upon the formation, growth and breakdown of the inhibition layer in the
galvanizing bath.
The purpose of this project is to assess the cratering formation mechanism on GA coatings
and its relation with IF steel reactivities whose main influence depend upon the segregation
of Ti and Nb compounds at the ferrite grain boundaries.
2. LITERATURE BACKGROUND
2.1. INFLUENCE OF ALLOYING ELEMENTS ON IF STEELS
The chemical elements Ti and Nb are very effective as microalloying elements in IF steels,
influencing the microstructure by both solute drag effect and the formation of nitrides and
carbides. Since the solubility product and the physical properties of each element and each
compound are different, there exist characteristic differences which cause each of these
elements to have specific merits.
Titanium forms nitrides, which are stable at high temperatures, and these Ti nitrides provide
control of the austenite grain size at the reheating temperature before hot working and also
in the welding, in particular in the heat affected zone close to the fusion boundary. The
elimination of free N due to the formation of TiN is positive for the toughness and indirectly
makes Nb more effective. Furthermore, the influence of Ti on sulphide shape control had
been widely used at a time, when the production of low S contents was not standard.
Niobium is reported by some authors as the most effective microalloying element for grain
refinement by controlling the austenite grain size during the reheating processes for heat
treatments like normalizing, quenching or carburizing, acting additionally to the traditional
AlN technology. Moreover, it has an outstanding status in retarding recrystallization during
austenite processing via thermomechanical rolling, resulting in grain refinement, which
cannot be obtained by any heat treatment process. Other outstanding effects of Nb, such
11
as lowering the austenite-ferrite transformation temperature by a solute drag effect of the
effective precipitation hardening potential can be used only to a certain extent due to its
limited solution in austenite. As a result, the combined usage of microalloying elements is
often the optimum solution.
2.2. C, N, S AND P STABILIZATION IN IF STEELS
Interstitial Free (IF) steels are ultra low carbon steels where all the interstitials, generally C
and N, have been completely combined with alloying elements and are characterized by
low yield strength and excellent formability i.e. high elongation and n-value even when
produced through continuous annealing and galvanizing lines.1
Therefore, Interstitial Free (IF) steel grain boundaries are essentially C free due to the
precipitation of Ti and Nb carbides within the grains.2
Regarding to Ti-stabilized IF steels, the more traditional approach is that Ti additions are
made to combine with N and S, while Nb additions are made to account for C, as shown by
the following relationships:
• Tistabilize = 3.42N + 1.5S;
• Nbstabilize = 7.74C.
According to this same traditional approach, the predominant precipitates would be TiN,
TiS, Ti4C2S2 and NbC, with the amount of each primarily depending upon composition.
Recent work has shown that Nb may also be present in a complex carbo-sulfide
(Ti,Nb)4C2S2.2
It has been noticed that, depending upon the Ti level, a substantial amount of Nb may
remain in solid solution. In such cases, the solute Nb may be beneficial for property
development, especially with respect to hot-dip galvanneal coatings.
Osman et al. 3 mentioned that Nb additions are typically made to stabilize interstitial
species. However, there are cases where solute Nb will be desired for property
development. Nb additions to these grades may be made to enhance other properties. This
12
solute Nb has been shown to segregate to grain boundaries, to subgrain boundaries, and to
free surfaces. Hook et al. apud Osman et al.3 showed that a strong relationship between
the solute Nb to the grain boundaries will become important in the microstructural
development of galvanneal coatings as well as cold work embrittlement resistance.
In these grades, interstitial stabilization primarily occurs through precipitation with Ti, similar
to that found in Ti-stabilized steels. The traditionally accepted precipitation sequence in Ti-
stabilized steels with atomic ratio S:C:N 1:1:1 is TiN, TiS, Ti4C2S2 and TiC, as the
temperature decreases. The amount of Ti required to fully stabilize IF steels was expressed
as (wt%): Tistabilize = 3.42N + 1.5S + 4C.3
Bhattacharya et al. 1 carried out studies on the behavior of Nb additions upon galvanneal
coatings. It has been stressed that Nb is also added to IF steels in order to produce ‘high
strength’ IF steels where Nb addition imports strengthening by grain refinement and by
precipitation hardening of NbC or Nb(CN). One effect of Nb, which has been well
established, is its effect on crystallographic texture. It is known that Nb additions improve
the r45 value and reduces ∆r compared to Ti-IF steels. The other well-known and interesting
effect of Nb is the improvement of the galvannealing behavior of IF steels. The work of
Tokunaga and Kato apud Bhattacharya et al. 1showed that the powdering tendency of Ti-
Nb-stabilized IF steels is superior to Ti-stabilized IF steels. Moreover, recent work by Cheng
apud Bhattacharya et al. 1 has shown that while under certain conditions, the galvannealing
rates of Ti-IF and Ti-Nb-IF steels are comparable, it is more consistent in Ti-Nb-IF than in
Ti-IF steels.
It is also worth noting that Nb concentration at the surface of annealed Ti-Nb-IF steel
sample is about three times of that for Ti-IF sample. Some GD-OES results clearly showed
that Nb segregates to the surface both at higher concentrations and to a higher depth in Ti-
Nb-IF steels. Through elemental profiles obtained by ESCA, below the outer layer of
contaminants (C and O), the segregation of Nb was easily detected in significant amounts
in Ti-Nb-IF steels and to a much lesser extent in Ti-IF steels.1
Feliu et al 4, in a study oriented to characterize oxides on the surface of galvanneal
coatings, observed that a slight increase in the number of white precipitates (Ti nitrides and
13
sulfides) was seen on the Ti-stabilized IF steels, compared with those observed on the dual
stabilized IF steel. However, they mentioned that there had been no correlation yet between
those precipitates and any relevant substrate differences.
2.3. MICROSTRUCTURE OF IF STEELS
The interstitial atoms such as C and N feature an important effect upon the deformation
behavior of steels. Indeed, discontinuous yielding in steels is strongly related to interstitial
species5,6.
In fact, it has been pointed out that a yield point is bound to stand still even with an
interstitial C content of 10 ppm. Furthermore, solute C and N may also minimize the
formability, especially work hardening, of steels7.
Moreover, discontinuous yielding can lead to Lüders bands formation that are detrimental of
surface aesthetics, and may also result in premature breakage during forming or even non-
uniform springback after deformation.
Therefore, in order to solve those detrimental effects of interstitial species as stated above,
Interstitial Free Steels have been developed whose microstructure is nearly free of solute C
and N.
Besides, these IF steels predominantly rely on the solid state precipitation of carbides,
nitrides, and carbo-sulfide especially in order to reduce the solute interstitial content as
shown on a previous session.
Additionally, the total C and N contents are reduced to extremely low levels, typically less
than 50 ppm wt C and less than 60 ppm wt N through the use of modern steelmaking
techniques at the steelmaking shop such as the circulation-type vacuum degassing
process3.
14
2.4. CONTINUOUS GALVANIZING AND GALVANNEALING
Hot-dip galvanizing is the immersion of a steel article in a liquid bath of Zn or a Zn alloy, by
batch or continuous process.8
Hot-dip galvanneal coatings are Fe-Zn alloy coatings, which typically contain between 8
and 12 wt% Fe.3 They are essentially diffusion coatings that expose the Zn galvanized steel
to an annealing temperature around 500 0C to produce a fully alloyed coating containing
Fe-Zn intermetallic phases. This is accomplished by inserting heating and cooling capacity
above the liquid Zn pot in order for the galvannealing process to be continuous.8
Since galvanneal is the process that converts pure Zn in coating into Fe-Zn intermetallic
compounds through diffusion between Zn and substrate steel, the phase evolution of
coating upon galvannealing strongly depends on the alloy layer formed first between molten
Zn and steel substrate.9
Galvanneal steel sheets are important for automotive and other applications because they
offer excellent corrosion resistance, good weldability, and good formability.10 Furthermore,
hot-dip galvanneal coating products are becoming particularly popular among car-makers
because of their moderate production costs, easy fabrication and good in-service
performances. Its sticking-free stamping capability, good spot weldability, and outstanding
perforation and cosmetic corrosion resistances are especially worth mentioning. In the case
of car body panels, particularly for exposed parts, coating adhesion is of prime
importance.11
Additionally, it is well know that galvannealed material offers a number of advantages over
galvanized sheet: better spot weldability and paint adhesion and less risk for sticking during
press forming.12
15
Usually an annealing thermal cycle is required prior to galvannealing. Many studies have
been published in which annealing cycles lasting up to several hours have been used to
reveal the segregation mechanisms of small alloying element contents towards the surface
of steels. These long annealing times are necessary to obtain a sufficient amount of
precipitates on the steel surface in order to characterize them by conventional techniques.
Far fewer researchers use annealing cycles that are similar (close to 1 minute) to those
commonly used during galvanizing in commercial lines.4
During galvanneal processing, it was observed that the development of the coating
microstructure can be divided into at least four distinct stages, as follows:
• Formation of the Fe-Al inhibition layer;
• Breakdown of the inhibition layer;
• Fe-Zn compound formation dominated by solidification of the coating by δ phase (Fe-
Zn intermetallic compound) formation;
• Fe enrichment of the coating via solid state interdiffusion of Fe and Zn.13
In the galvannealing process, the cooling tower is transformed into an annealing furnace
either by introducing an induction furnace above the gas wiping jets and using a good part
of the tower height as a holding furnace, or introducing a gas heating furnace above the gas
knives.
On the one hand, bath temperature and dipping time, a result of line speed, are specific
operating parameters that can be optimized to control efficient surface cleaning, strip
heating and minimization of alloy layer growth during the immersion step.
On the other hand, in order to produce an optimized galvanneal product, it is important to
control processing parameters such as heating rate, peak or hold temperature, and cooling
rate, as they all affect the amount of alloying that will occur in the coating prior to reaching
the tower roll.
Marder et al.8 affirmed that good process control requires that the effects of heating rate,
hold temperature and time, and cooling rate on the Fe-Zn reaction kinetics be well
understood so that the optimum coating for the desired properties can be obtained. Besides
16
processing variables, variations in bath chemistry and substrate composition all contribute
to the final microstructure.
Besides the variables associated with hot-dip galvanizing, e.g. bath temperature, substrate
alloy content and bath composition, a major variable of the galvannealing process is the line
speed, which will control the dependent galvannealing variables such as:
• Heating rate;
• Peak or holding temperature and time;
• Cooling rate.8
Furthermore, the water vapor content in the annealing atmosphere was expressed as the
dew point value. For IF steels, Feliu et al.4 reported that the annealing cycle consisted of a
linear heating ramp (15 0C/s) up to 8500C, holding the specimens at this temperature for 40
s and then immediately cooling them at a rate of 15 0C/s to room temperature.
2.4.1. FORMATION OF GALVANNEALED COATINGS
2.4.1.1. THE Fe-Al INHIBITION LAYER
During galvanneal processing, it was observed that the development of the coating
microstructure can be divided into at least four distinct stages, as follows:
• Formation of the Fe-Al inhibition layer;
• Breakdown of the inhibition layer;
• Solidification of the coating by δ formation;
• Growth of Fe-rich phases via solid state interdiffusion of Fe and Zn.14
The inhibition layer was observed to be a layered structure of FeAl3 and Fe2Al5. Other
authors have found the layer to consist of Fe2Al5, FeAl3 and/or an Fe-Al-Zn ternary
compounds.15 The formation and dissolution of the inhibition layer is an important issue
because it sets the stage for the coating transformation that follows.14
Dionne et al.16 postulated that the formation of an inhibition layer is dependent upon the
ferrite grain substrate. They found that the regions associated with the emerging ferrite
17
grain boundaries showed deeper attack with no residual Al-rich layer, whereas a thin Al-rich
layer was detected on the surrounding surfaces where the reaction was less advanced.16
Results achieved by McDevitt et al.17 through TEM analysis show that the inhibition layer
displayed two distinct morphologies each covering regions approximately 50 µm2 in size.
The first region was a flat continuous layer, and the second region was highly discontinuous
where only 60% of the substrate was covered. Furthermore, it was observed that the
structure of the inhibition layer was not strongly influenced by the chemical composition of
the substrate steel.
Their TEM observations on three different steel substrates can be summarized as follows:
• After 1s of dipping, an Fe-Al inhibition layer formed on all three substrates and could
consist of Fe2Al5, FeAl or and Fe-Al solid solution. On top of that layer, Γ1, ζ or η-Zn
was observed.
• After 3s, the dominant microstructure was a thin Fe2Al5 inhibition layer covered by a
50-150 nm thick Γ1 layer, and occasional ζ crystals.
Additionally, electron diffraction on cross-sectional TEM samples identified the Al-rich layer
as a structure consisting of Fe2Al5 at the steel interface and Γ1 and ζ lying between the
Fe2Al5 and the Zn outer layer. The presence of Γ1 and ζ intermetallic compounds on top of
the Fe-Al inhibition layer is quite surprising. It has been a widely held belief that the Fe-Al
layer prohibits the formation of any such Fe-Zn compounds.17
Some studies reported that Fe2Al5 is the main constituent of the Al-rich layer when high Al
(> 0.13 wt%) baths are used. However, more recent studies reported that the Al-rich
interfacial layer is composed of two Fe-Al intermetallic phases: Fe2Al5 and FeAl3. The
sublayer in contact with the Zn overlay is mainly FeAl3, while the Al-rich layer on Ti-IF and
Ti-Nb-IF steels was composed primarily of Fe2Al5. The preferred orientation relationship
between Fe2Al5 and the ferrite grains of the substrate was identified as (011)Fe//(001)Fe2Al5,
[-100]Fe//[-310]Fe2Al5. It has also been reported that only 0.09 wt% Al is required to slow
down the Fe-Zn reaction.10
18
Morimoto et al.18 confirmed that the formation of an Fe2Al5 interfacial layer during
galvanizing is the most widely accepted result of many experiments and had been reported
by several authors. However, there are also reports apud Morimoto et al.18 that the
interfacial layer is composed of Fe-Al-Zn ternary compounds, mixtures of Fe2Al5 and FeAl3,
and some researchers Morimoto et al.18 report observing Fe-Zn compounds instead of an
Fe-Al interfacial layer.
They also stated that the Fe2Al5 was the primary phase in the interfacial layer and formed a
continuous layer on the steel substrate. The Fe2Al5 grains could frequently be recognized
by characteristic triangular-shaped grain cross-sections. A second Fe-Al intermetallic
phase, FeAl3, was observed to lie between the Fe2Al5 and the Zn layer. However, the FeAl3
was not ubiquitous in the coating. The thickness of the layer of Fe2Al5 and FeAl3 was
confirmed to vary between 50 and 300 nm.18
Furthermore, at the grain boundary, the Zn composition was nearly twice that of the bulk
grain due to the Zn enrichment at the grain boundary observed through a compositional
profile across a grain boundary between two Fe2Al5 grains. There was a variation in Zn
content from grain to grain in Fe2Al5. 18
According to Lin and Meshii, the Fe2Al5 phase shows some pillar-like grains of Fe-Al-Zn
ternary compound. On an ELC (<0.010 wt% C) steel substrate, they observed only the Fe-
Al-Zn ternary compound.18
Regarding to selective competition between the formation of Fe-Zn intermetallic compounds
and the inhibition layer, O’Dell et al. have observed that ζ phase can form epitaxially upon
(111) orientated grains and a melting point of 5300C. The ordered epitaxial nucleation of the
ζ phase upon the substrate becomes energetically favorable due to the reduction in the
energy barrier as a result of easy accommodation upon (111) oriented grains. Hence, the
Fe-Zn intermetallic compound formation is energetically favorable over the Fe2Al5 inhibition
layer.19
19
In order to prepare samples for inhibition layer characterization, many researchers have
done so through selective dissolution of the Zn overlay by using fuming nitric acid (e.g. 90
vol% HNO3 according to Baril et al.10). The nitric acid dissolved Zn and Zn-Fe compounds,
leaving the Fe-Al phase untouched at the surface of the steel substrate.10
Regarding to microscopy analysis, it is important to mention that the Fe-Al layers that form
in commercial coatings are generally too thin to be observed in cross-section using light
optical microscopy or conventional SEM. Therefore, high intensity X-ray sources or parallel
beam geometry diffractometers are necessary to detect diffraction peaks from the interfacial
layer.18
TEM allows for superior spatial resolution and unambiguous phase identification using
electron diffraction. SEM provides a broader view of the sample and a clear image of the
morphology of the interfacial layer. X-ray diffraction offers the advantage of study from a
statistically significant area of the sample. However, the results can be difficult to analyze
due to the complexity of the phases in the Fe-Al-Zn system.18
With the exception of Lin et al. apud McDevitt et al.15, who used electron diffraction in TEM,
researchers have used either X-ray diffraction (XRD) exclusively or in combination with
EPMA to identify the thin inhibition layer. McDevitt et al.15 commented that there exists
some ambiguity in results carried out by both XRD and EPMA analysis because:
• The XRD patterns are very complex and some authors did not precisely identify each
diffraction peak;
• There is still some debate regarding the lattice parameters, and thus the correct
location of diffraction peaks for Fe2Al5;
• The thickness of the inhibition layer can be small, such that the diffracted intensity
from the layer is very weak;
• EPMA identification of a phase is only reliable if the layer has reached an equilibrium
state. Supersaturation of one element in a phase can lead to incorrect identification
of the phase.15
2.4.1.2. FORMATION OF THE Fe-Al INHIBITION LAYER
20
There are often conflicting results about the formation of initial inhibition layer. It is believed
that many of the inconsistencies made in observations are due to the nonequilibrium nature
of this process. In fact, it is very difficult to explain the decomposition of this inhibition layer
because the Fe2Al5 is thermodynamically stable in the presence of liquid Zn. Many of the
inconsistencies observed are believed to be caused by the differences between
nonequilibrium and equilibrium conditions.20
It has become clear that the strict use of equilibrium arguments fails to describe the
microstructural development observed in galvanneal coatings, especially early in the
processing where the local chemical environment might deviate significantly from
equilibrium, i.e., occurrence of Fe supersaturation in the Zn bath.14
Nevertheless, Tang showed that the formation of the inhibition layer is a two-stage process.
The first stage, associated with a high rate of Al uptake at the coating/substrate interface, is
controlled by continuous nucleation of Fe2Al5, followed by a second stage diffusion-
controlled growth process.8 Additionally, he suggested that the inhibition layer formed on
the steel surface by precipitating from an Fe-supersaturated Zn liquid. This mechanism
would therefore require the coprecipitation of FeAl and Fe2Al5 from the bath.
McDevitt et al. studied the steps of inhibition layer formation, growth and breakdown for
three different steel substrates. In the first second of hot-dipping the majority microstructure
on all three steel substrates consisted of an Fe-Al layer on the steel surface covered by
layers of Γ1 or ζ phases. Moreover, the inhibition layer was absent on approximately 15% of
the steel surface after dipping for one second, in which case the substrate was observed in
direct contact with either Γ1 or ζ crystals. Furthermore, Fe2Al5 appeared to comprise
approximately 85% of the inhibition layer.14
O’Dell et al. reported that the rapid formation of the inhibition layer, about 5 miliseconds, is
due to Al’s high affinity with Fe and also due to it small critical radius.19
2.4.1.3. GROWTH OF Fe-Al INHIBITION LAYER
21
McDevitt et al. reported that, after three seconds, the inhibition layer grew laterally on the
substrate surface so that it covered at least 95% of the surface. At this point only Fe2Al5
was observed in the inhibition layer, and no Al was detected in the substrate adjacent to the
Fe2Al5 layer.
Moreover, it appeared that substrate chemistry had little or no substantial influence on how
the Fe2Al5 layer developed. The only slight substrate effect observed was that the ζ phase
appeared more uniformly distributed on the P-added substrate than on the Ti-Nb IF or Ti IF
substrates.14
Morimoto et al. added that, whenever the two phases Fe2Al5 and FeAl3 coexist, Fe2Al5 was
always observed in contact with the steel substrate. It is likely that the microstructure
developed by FeAl3 formed first on the steel substrate, followed by subsequent precipitation
and growth of Fe2Al5. There would be a thermodynamic driving force to nucleate Fe2Al5 at
the interface between FeAl3 layer and the Fe substrate. Once nucleated, Fe2Al5 could grow
by consuming FeAl3 as Fe diffused into the layer. At the end of the day, one might say that
Fe2Al5 is responsible for inhibiting the growth of Fe-Zn compounds.18
2.4.1.4. BREAKDOWN OF Fe-Al INHIBITION LAYER
Dionne et al. observed that the first step of the breakdown of the inhibition layer at the
beginning of the galvannealing reaction is the enrichment of the inhibition layer with Zn. The
second step in the breakdown of the inhibition layer is the transformation of the Zn-enriched
Fe2Al5 phase into a layer of δ crystals.16
It has been thought by many authors that the mechanism that operates in the inhibition
layer breakdown may be strongly influenced by the initial state of the compound layers.14
Marder et al. stated that the inhibition layer is a transient phenomenon in galvannealing
coatings, explained by thermodynamic equilibrium between Fe2Al5 and the liquid. If some
regions of overlay/substrate interface have little Al content, Zn atoms attack substrate
through ferrite grain boundaries, which are short circuit paths for Zn atom diffusion. When
Zn reaches the substrate, it reacts with Fe, nucleating Fe-Zn intermetallic phases at the
22
Fe2Al5 inhibition layer/substrate interface that bursts the layer apart into the surrounding
bath. Since Zn atom diffusivity is too high, it diffuses back to the overlay forming outbursts,
totally destroying the Fe-Al inhibition layer.8
Similarly, McDevitt et al. observed that substrate chemistry had little or no effect on the
formation of the inhibition layer, but strongly affected the breakdown of the Fe-Al layer
where the inhibition layer survived significantly longer on the P-added steel substrate.
There was no evidence that the inhibition layer disappeared due to mechanical breakage by
growth of Fe-Zn compounds in outbursts between the layer and the steel substrate on
samples galvannealed at 500 0C.14
Therefore, they stated that the inhibition layer breakdown did not occur via outburst
mechanism, rather it was eliminated by dissolution into the substrate and possibly by
consumption by the Γ1 phase. It was reported that the Fe2Al5 inhibition layer began to
disappear upon removal from the Zn bath and heating to 500 0C. The layer was nearly
completely dissolved on the Ti-Nb IF and the Ti IF substrates by the time the sample
reached 500 0C.14
From three key observations regarding breakdown of the inhibition layer, McDevitt et al.
proposed their theory:
The thickness of the inhibition layer decreased with processing time;
The composition of the inhibition layer changed;
The composition of the substrate originally in contact with the inhibition layer changed.14
Thus, it was assumed that there is direct evidence that the inhibition layer dissolved at least
in part by Al diffusion into the steel substrate, according to the following statements:
The thickness of the inhibition layer decreased with processing time. The Al concentration
in the substrate near the coating/substrate interface was observed to increase when the
inhibition layer had disappeared, and the Zn concentration of the inhibition layer increased
with processing time. These observations support a dissolution mechanism where the
inhibition layer is dissolving by Al diffusing into the substrate at the substrate/inhibition layer
interface.14
23
Al is expected to diffuse into the substrate as the chemical potential of Al in Fe is lower than
in Fe2Al5 until the Al concentration of the Fe exceeds 17 wt%. However, due to the
thickness of the substrate relative to that of the Fe2Al5 layer, there can not be enough Al
available for the substrate to reach the equilibrium concentration of Al. There will always be
an Al concentration gradient in the substrate such that Al will diffuse away from the Fe2Al5
inhibition layer, and the Fe2Al5 layer will continue to dissolve.14
Zn enrichment of the Fe2Al5 concurrent with the Al depletion is not unexpected as the
system seeks an equilibrium condition. The combination of Al depletion and Zn enrichment
lead to the disappearance of the inhibition layer by dissolution into the steel substrate and
possibly simultaneous consumption of the inhibition layer by the existing Γ1 phase.14
Finally, on non-P added substrates, a higher Al concentration was observed at Fe substrate
grain boundaries. This observation indicates grain boundary and possibly interfacial
diffusion, which are significantly faster than bulk diffusion, are operating in the present case
of diffusion of Al into the substrate.14
Morimoto et al. also states that it is not likely that substrate grain boundaries play an
important role in the breakdown of the inhibition layer, since no evidence that grain
boundaries in the steel substrate were locations of enhanced formation of the inhibition
layer, nor were the grain boundaries a site for the local breakdown of the inhibition layer.
Substrate grain boundaries did not show penetration of Al or Zn indicating that their
diffusion into the steel substrate is also not likely an important factor in the breakdown of
the inhibition layer.18
In fact, the composition profile in the direction normal to the interface indicates that bulk
diffusion of Zn into the inhibition layer may not be an important diffusion mechanism
because the Zn concentration was unchanging with distance into the Fe2Al5 grain. In
contrast, the profile across the Fe2Al5 grain boundary indicating that grain boundary
diffusion of Zn through the inhibition layer may be important in the breakdown of this layer.18
24
On the contrary, Jordan et al. observed that the effect of substrate grain size on Fe-Zn
phase formation and growth supports the mechanism that the Fe-Al inhibition layer is first
attacked by liquid Zn at sites corresponding to substrate steel grain boundaries. This theory
will be discussed in a further session where the outburst mechanism is more clearly
explained. 2
Finally, Guttman proposes a diffusion-assisted mechanism to explain the inhibition
breakdown and the localization of the outburst. Indeed, the diffusion of Zn towards the steel
substrate through the Fe2Al5 layer could be the mechanism that explains the destruction of
the diffusion barrier, which is called the inhibition layer. The diffusion short circuits through
Fe2Al5 may be the grain boundaries of the layer. 10
2.2.1.5. EFFECT OF OPERATIONAL PARAMETERS ON THE INHIBITION LAYER
McDermid et al. reported that bath effective Al controls the formation of the η-Fe2Al5Znx
layer at the steel/coating surface such that, at low bath Al contents, the layer is thin and
inhibition layer occurs rapidly, facilitating the galvannealing operation. On the other hand, at
high bath Al contents, it promotes the formation of a thicker inhibition layer in which
inhibition breakdown is prevented and a pure metallic coating results.21
Marder et al. reported that increasing the Al content or decreasing the bath temperature
increases the stability of the Fe-Al inhibition layer. The stability of the Fe2Al5 interfacial layer
is enhanced by high Al levels (> 0.15 wt% Al) inhibiting Fe-Zn phase formation in the short
term, whereas Fe-Zn reactions are enhanced by low Al bath levels (< 0.15 wt% Al).8
Many researchers believe that there is a threshold bath Al level around 0.14 wt% below
which no Fe-Al inhibition layer should form. However, Pelayo et al. presented results that
conclusively show that an Fe-Al interfacial layer forms under the low bath Al content (0.14
wt% total Al), short dipping time (3 seconds) conditions of commercial hot dip
galvannealing. Additionally, they observed that, as the dipping time in the Zn bath
increased, the coverage of the substrate by the Al-rich layer increased.
25
Moreover, the surface of the Fe-Al interfacial layer was generally flat and compact. In
contrast, the surface morphology of the inhibition layer formed in a 0.20 wt% Al bath was
sharply faceted. Some of the differences in morphology and microstructure of the grains of
the inhibition layer in coatings produced in a 0.14 wt% Al and 0.20 wt% Al baths may be a
result of FeAl3 being present on the surface for one processing condition and Fe2Al5 being
present on the surface in the other processing condition. The difference in surface
morphology may also result from different growth mechanisms of the Fe-Al layer in the low
and high Al content Zn bath processing conditions.15
Morimoto et al. observed that, in long dipping experiments, the supply of Al to the interface
is effectively unlimited and controlled only by the rate at which Al can diffuse within the
molten Zn bath. In commercial production, the amount of Al is constrained as soon as the
trip exits the bath to that present in any interfacial layer plus that in the liquid Zn layer
exiting with the strip. This difference may have an important effect on the growth or
subsequent dissolution of the interfacial layer.18
Baril et al. stated that Al content in the inhibition layer generally increases with immersion
time and increased Al content in the bath. Furthermore, longer immersion times in the Al-
containing Zn lead to a more complete Al-rich layer and thereby reduce the amount of Fe in
the coating as a result of Fe-Zn reactions during cooling. Yet this is not uniform. Hence,
longer immersion times also lead to an increase of the coating Fe, due to Fe-Zn reactions in
some areas, while the Al-rich layer grows in other areas.
On the other hand, when the Al level in the bath is below 0.12%, the Fe/Al ratio is strongly
related to the immersion time, which suggests that the rate of the Fe-Al reaction is lower
than that which occurs for the intermetallic formation of Fe-Zn. Thus, for this relatively low
Al content (0.12%), the Fe-Zn reactions and the formation of Al-rich layer take place
concurrently.
They added that one major effect of rapid solidification and increased metastability is an
extension of solid solubility limits. Consequently, more excess Zn can be present in the
Fe2Al5, and the Zn content of the Al-rich layer should be between the ZnFeAl3 phase and
the stable Fe2Al5 phase. By its way, the Fe2Al5 crystals have a strong crystallographic
26
texture. X-ray diffraction using parallel beam optics showed a strong (513) reflection. This
reflection had been related to (200) planes parallel to the surface of the substrate. In
addition, this crystallographic texture of the Al-rich layer is related to the Al content of the
bath. The interfacial layer studied was mainly constituted by Fe2Al5 with (200) planes
parallel to the substrate surface.10
Reporting results related to the effect of substrate grain size upon the inhibition layer,
Jordan et al. observed that substrate grain size is a more significant factor in Al-containing
baths than in the 0.00 wt% Al-Zn baths, in which the Fe2Al5 inhibition layer does not form,
negating any incubation time necessary for Fe-Zn phase nucleation and growth. A large
grain seized ULC substrate promoted the stability the Fe-Al-Zn inhibition layer and delayed
the formation of Fe-Zn phases.2
Faderl et al. reported that increasing the difference between the strip entry temperature and
the Zn bath temperature increased the Al content of the interfacial layer.15
2.2.1.6. EFFECT OF SUBSTRATE CHEMISTRY ON THE INHIBITION LAYER
In order to determine how the different substrate compositions influence the beginning of
galvannealing reaction, interrupted galvannealing treatments are performed to induce
partial breakdown of the inhibition layer. 16
Miyasaka et al. concluded that the acceleration or retardation of galvannealing reaction by
the addition of C, P, Ti, Nb or B alloying elements was not caused through the change of
the amount and/or structure of Fe-Al-Zn initial layer.22
Morimoto et al. observed that the grain size of the interfacial layer on the IF substrate
sample was significantly larger (0.5 µm) in the lateral dimensions than on the P, Mn-added
sample (0.2 µm), but both samples showed the same faceted morphology. Moreover, the
grain size was the only difference observed between the two samples. The composition,
thickness, and the constituent phases of the interfacial layer were the same for both
substrates.
27
Nevertheless, the difference in grain size between Fe2Al5 formed on the two substrates
suggests that the segregation of P and Mn on the substrate surface decreases the
nucleation rate of the interfacial layer and delays the formation and growth of the inhibition
layer. This result may be related to observations that the rate of the galvanneal reaction is
slower on P-added steel substrates than on IF substrates.18
Really, Lin et al. apud Jordan et al.23 proposed an alternative mechanism, whereby P
segregates to the steel surface during recrystallization annealing, stabilizing the inhibition
layer and retarding the rate of Fe-Zn phase growth reactions during galvanizing.
However, Jordan et al.23 observed that, in a 0.00 wt% Al-Zn bath, the kinetics of the total
alloy layer and individual Fe-Zn phase layer growth were not influenced by the ion
implantation of P. The total layer growth as well as the individual phase layer growth
followed the same growth-time relationships on both the P- and non-P-ion implanted
surfaces. Coverage of the steel/coating interface by the total Fe-Zn alloy layer was
complete on both surfaces, indicating the P-ion implantation did not retard or inhibit the
overall Zn attack of the steel surface. Briefly, they concluded that P-ion implantation did not
affect the composition of the Fe-Zn-Al reaction layer. However, the total thickness of the
Fe2Al5 (Zn) layer was found to be slightly larger at all reaction times for the non-P-ion
implanted surface.
Similarly, in a 0.20 wt% Al-Zn bath, growth kinetics of the Fe2Al5(Zn) showed a linear
growth rate (n=1) for both surfaces, independent of P surface segregation. It appeared for
them that the grain size of the substrate and its effect on the grain size of the Fe-Al-Zn
inhibition layer are the dominant reaction mechanisms as compared to surface chemistry (P
surface segregation).
Guttman et al. apud Jordan et al.23 proposed a mechanism that liquid Zn diffusion is the
most rapid at random high-angle boundaries of the Fe2Al5 inhibition layer, and these high-
angle boundaries could correspond in location to the sites of underlying substrate steel
grain boundaries. Thus, in contradiction to the suggestion by Lin et al., P segregation to
substrate steel surfaces is not an important mechanism in the Fe-Zn reaction. Instead,
substrate grain size dominates the nucleation and formation of Fe-Zn phases with P only
28
acting to retard Zn diffusion along substrate steel grain boundaries by segregating in these
boundaries, in agreement with Allegra et al. apud Jordan et al.23
Lee et al. studied the influence of Ni additions upon the inhibition layer. They observed that,
in the absence of Ni addition, Al is mainly concentrated at the interface between Zn coating
and the base steel. As a result, a dense Fe-Al inhibition layer forms at the interface
presumably acting as a diffusion barrier. In contrast, the Al concentration at the interface
was remarkably reduced by the 0.15 wt% Ni addition, and the Al distribution over the
interface is almost the same as that of Ni. This means that Ni strongly interacts with Al at
the interface. Therefore, it seems that the inhibition layer of Fe2Al5 decomposes thermally
and a new phase of Al-Ni intermetallics forms in the coating layer. These intermetallics
presumably accelerate the Fe-Zn alloy reaction.20
Finally, Dionne et al. carried out Mn and Ti mapping of FIB cross-sections. They revealed
numerous surface oxides located within the inhibition layer. The surface oxides did not
affect the thickness of composition of surrounding inhibition layer phases and the Fe2Al5
phase was not alloyed with either Mn or Ti. 16
2.4.1.7. EFFECT OF IF STEEL ALLOYING ELEMENTS ON THE INHIBITION LAYER
Bhattacharya et al. observed that Nb segregation to the surface may affect galvannealing in
two ways. It may affect the diffusion of Fe into the Zn layer during galvannealing or it may
affect the formation of the inhibition layer that, in turn, will influence the galvannealing
process. In any case, their study proved that Nb segregation to the surface, in addition to
grain boundary segregation, explains why the galvannealing behavior of Ti-Nb-stabilized IF
steels is markedly different than Ti-stabilized IF steels.1
Similarly, Osman et al. reported that the breakdown of the inhibition layer for Ti-stabilized
steels occurred randomly, but was more uniform for Ti-Nb-stabilized steels.3
However, Miyasaka et al considered that C, Ti and Nb did not affect the amounts and
structure of Fe-Al-Zn barrier layer, but the influence of Ti and Nb was greater on the
reduction of Ceff – whose formula usually is Ceff = C – (Ti – N.48/14)/4 – Nb.12/93 – and
29
presumably C segregated at the ferrite grain boundary, and resultant change in metal
diffusion through grain boundary, than on the stability and protectability of the Fe-Al-Zn
barrier layer.22
Meshii et al. supported the theory that, since the growth of the δ phase did not occur until
Fe2Al5 layer had decomposed, it is suspected that the non-uniform growth can be
contributed in part to non-uniform dissolution of the inhibition layer on the Ti IF and Ti-Nb IF
substrates. Such behavior is expected because substrate grain boundaries acted as sites
for enhanced Al dissolution from the Fe2Al5 into the substrate steel which should result in
faster dissolution of the Fe2Al5 layer at the substrate grain boundaries.13
Dionne et al.16 reported that the major phase in the inhibition layer was Fe2Al5 containing 7-
11 wt% Zn for a Ti-stabilized IF steel and 5-14 wt% Zn for a Ti-Nb-stabilized IF steel.
Moreover, for the Ti IF steel, a Fe-Al-Zn phase with 40-50 at% Al and up to 30 at% Zn was
observed, suggesting that significant diffusion of Zn had taken place during galvanizing
and/or during post-galvanizing cooling.
In previous studies, the FeAl3 phase had been detected as a minor constituent of the
inhibition layer on Ti-Nb and Ti-Nb-P IF steels galvanized in a hot dip simulator and in a
continuous galvanizing line. A small amount of FeAl3 in the inhibition layer would be unlikely
to affect significantly the galvannealing response.
The size of the inhibition layer grains is of interest since it could have a significant influence
on the kinetics of inhibition breakdown during galvannealing. In this respect, the Ti IF steel
showed the thinnest minimum inhibition layer, of 30 nm, and the Ti-Nb IF steel a thickness
of 45-50 nm.
Moreover, the breakdown of the Fe2Al5 layer in the case of the Ti- and Ti-Nb-IF steels was
significantly faster in the regions surrounding the emerging ferrite grain boundaries than in
the bulk of the grains.16
Pelayo et al.15 studied the inhibition layer upon three different steel substrates. They carried
out SEM analysis that convincingly demonstrate that a nearly continuous, Al-rich interfacial
30
layer forms and was present on a Ti IF, Ti-Nb IF, and P-added Ti IF steel substrates after
dipping in a 0.14 wt% Al-bearing Zn bath for 3s.
Furthermore, the Fe-Al interfacial layer on all three substrates (Ti, Ti-Nb and P-added IF
steels) was identified using electron diffraction as being composed primarily of Fe2Al5.
Moreover, the Fe2Al5 layer had a nominal thickness of 30 nm on the Ti-Nb IF substrate and
10-20 nm on the P-added and Ti IF substrates.
They also observed that the Ti IF substrate exhibited the most continuous inhibition layer of
the three substrates with 90–95% coverage of the substrate. The Ti-Nb IF substrate
showed a compact Fe-Al layer covering approximately 80% of the substrate surface.
Finally, they concluded stating that their results showed an inhibition layer where Fe2Al5 is
the dominant phase. Hot-dip galvanized samples were composed primarily of Fe2Al5, but
minor amounts of FeAl3 were also present in the layer. Moreover, in the early stages of
development, the inhibition layer on the Ti and Ti-Nb IF substrates went through a transition
from an Fe-Al solid solution or ordered phase to the Fe2Al5 phase.15
2.4.2. GROWTH OF Fe-Zn INTERMETALLIC COMPOUNDS
A reasonable understanding of the galvanneal process and final microstructure is strongly
dependent upon the nucleation and kinetics of Fe-Zn intermetallic compounds formed in
between the Zn layer and the steel substrate. Indeed, it is the formation and growth of the
Fe-Zn intermetallic compound layers that ultimately controls the properties of galvanneal
coatings.13
Table 1 shows the consensus among most part of the authors regarding to some features
of the of Fe-Zn intermetallic phases:8
Table 1. Features of Fe-Zn intermetallic compounds.
Intermetallic Phase Fe (wt%) Symbol Crystal Morphology
ζ (Zeta) 5.7-6.3 FeZn13 Monoclinic Pillar-Like
31
δ (Delta) 7-11.5 FeZn10 HCP Columnar
δ1 (Delta 1) 7.5-12 FeZn7 HCP Polygonal
Γ (Gamma) 23.5-28 Fe3Zn10 BCC Columnar
Γ1 (Gamma 1) 17-19.5 Fe5Zn21 FCC -
Regarding to the usual chemistry range of Fe-Zn intermetallic compounds, Perrot et al.
introduced the data shown in Table 2.21
Table 2. Usual chemistry range of Fe-Zn intermetallic compounds.
Intermetallic Phase wt% Al wt% Fe wt% Zn Fe/Al ratio
η-Fe2Al5Znx 37-46 31-37 18-25 0.7-0.9
δ-FeZn7 1.5-3.5 2.2-9.5 87-93 1.55-5.30
ζ-FeZn13 0.7-1.0 5.8-6.1 93.2 N/A
According to Marder et al.8, the zeta (ζ) phase is formed from the peritectic reaction
between the delta (δ) phase and liquid Zn at 530 +/- 10 0C. The delta (δ) phase is formed
from another peritectic reaction, gamma (Γ) and liquid, at 665 0C. The gamma1 (Γ1) phase
forms as a result of a peritectoid reaction between the gamma (Γ) and delta (δ) phase at
550 +/- 10 0C. The gamma (Γ) phase forms as a result of peritectic reaction at 782 0C
between ferrite and liquid Zn.
Additionally, they classified the cross-section microstructure of galvanneal coatings as
follows:
Type 0: Underalloyed coating containing predominantly ζ phase;
Type 1: Optimum alloyed coating with less than a 1 µm interfacial Γ layer and an overlay
containing δ phase interspersed with a small amount of ζ phase.
Type 2: Overalloyed coating with a Γ layer > 1 µm and an overlay containing a δ phase with
basal plane cracks perpendicular to the coating/substrate interface and an occasional top
layer of ζ phase.
Moreover, the relative amounts of ζ, δ and Fe2Al5 phases depend upon:
32
• The substrate steel reactivity;
• Bath temperature;
• Immersion time.
Marder et al.8 reported that the sequential nucleation of Fe-Zn phases occurs at the
interface beginning with ζ phase layer, followed by δ phase layer, and after some incubation
time, Γ phase layer. δ and Γ are the brittle phases in the Fe-Zn coatings, while ζ phase is
the most ductile.
If the Zn melt is supersaturated with Fe, and there is sufficient nucleation of new crystals,
numerous tiny ζ crystals can form in the melt, that are separated from each other by the
solidified Zn η phase. It means that the morphology of ζ phase is a function of Fe
supersaturation on Zn bath.8
Baril et al.10 stated that ζ (FeZn13) crystals are preferentially formed orderly on (111)αFe,
whereas the precipitation of Fe-Zn intermetallics is retarded on (001)αFe and (101)αFe.10
Lin et al. 9 reported that the δ phase is the major phase in the commercial galvanneal
coatings with the Fe content in the range of 10 to 12 wt%.
They also observed a two layer structure, one consisting of the δ phase with a small fraction
of the ζ phase dispersed on the surface and Γ phases and another consisting of the δ and
Γ1 phases, was observed in the P-added IF steel sample substrate. On the other hand, a
three-layer structure consisting of the δ, Γ1 + δ, and Γ phases was observed on the IF steel
sample substrate.
Referring to a controversy in earlier diffraction studies, they mentioned that it was wrongly
speculated that the Γ1 phase is not likely to form in commercial galvanized steels because
of their short processing time. Finally, they stated that the existence of the Γ1 phase in
commercial galvanneal coating was then verified. The grain morphology of the Γ1 phase
was rather irregular in comparison with that of the Γ phase, which always exhibited a well-
defined columnar grain microstructure.
33
The Γ phase, which shows a columnar grain structure, has an average columnar width of
0.25 µm, which is a small fraction of the average grain size of steel substrate. This suggests
that no unique orientation relationship exists between the Γ phase and ferrite.9
According to some authors, e.g. Perrot et al, there are strong indications that the
galvanizing bath is not at chemical equilibrium. Hence, the metastable phase diagram and
not the chemical equilibrium is one of the most use to galvanizers for the calculation of
effective Al (by displaying only δ-FeZn7 and η-Fe2Al5Znx phases, rather than the higher
number of intermetallic species found over the same composition range showed in the
equilibrium diagram).21
Finally, among the microscopy techniques used to identify the Fe-Zn intermetallic
compounds, the electron probe microanalysis (EPMA) has been used to identify the phases
revealed by chemical etching. The EPMA analyzer, whose best estimated spatial resolution
is about 1 µm, is often unreliable in identifying a phase whose size is lesser than the spatial
resolution of the analyzer (e.g. Γ phase with thickness lesser than 1 µm). The X-ray
diffraction technique is frequently used to identify the phases in galvanneal coatings, but it
cannot be used to determine their distribution neither to detect a phase with a small volume
fraction.
The phase identification problem is especially pronounced for state-of-the-art high-tech-
processed galvanneal steel coatings, since the coating thicknesses are dramatically
reduced as compared to those produced during the early developmental stages of the
galvanneal-processing technique.9
2.4.2.1. KINETICS OF Fe-Zn INTERMETALLIC COMPOUNDS
Marder et al.8 reported that the incubation period (i.e. the time for Fe-Zn phases to form)
increases with an increase in Al content in the bath and decreasing bath temperature.8 Yet
Miyasaka et al.22 described the incubation time as the time between the start of heating and
the onset of rapid increase in the emittance. Thus, emittance is strongly related to that time
34
when there is a change of the coating surface from the liquid Zn into the solid Zn-Fe
crystalline phases, such as ζ and δ1 phases.22
Studying kinetics of Fe-Zn intermetallic compounds becomes a complex issue because
several interface reactions are occurring at the same time, including wetting of the solid
substrate by liquid Zn, dissolution of the steel by the Zn, isothermal solidification of Fe-Al-Zn
intermetallic compounds, solid state diffusional phase transformations, and solidification of
the liquid Zn alloy.
Nevertheless, the former researcher proposed a time sequence model for galvanneal
coating formation, as follows:
t0: The steel develops an Fe-Al interfacial layer during hot-dip galvanizing which, depending
on Al content in the Zn bath, inhibits the formation of Fe-Zn phases. The effectiveness of
the inhibition layer depend on the Al content in the Zn bath, as well as time of immersion
and bath temperature.
t1: Fe-Al inhibition layer breaks down during annealing causing nucleation and growth of δ
phase at the coating/surface interface. Outburst formation can also occur with accelerated
growth of Fe-Zn phases. Simulated annealing of a galvanized coating in an environmental
SEM shows the formation of outbursts at the interface. The rate of Zn attack will depend on
the substrate alloy addition as well as the temperature profile in the annealing process (type
0 galvanneal microstructure).
t2: As annealing continues, diffusional growth of δ phase in a columnar growth morphology
occurs. Previously nucleated ζ phase transforms to δ phase. Additional ζ phase may
nucleate due to oversaturation of Fe in the liquid η phase or upon cooling. An interfacial Γ
phase forms at the coating/steel interface.
t3: Zn depletion occurs with longer temperature exposure resulting in the complete
consumption of η phase at the coating surface. Fe concentration increases in the coating
as δ phase continues to form, pushing the ζ phase to the surface and maintaining a
constant 1 µm thickness Γ phase. This has led many to rely on an Fe concentration
measurement (approximately 10 wt%) as a requirement for the optimum galvanneal coating
(type 1 galvanneal microstructure).
35
t4: With longer times at temperature, δ phase diffusional growth continues towards the
surface of the coating consuming the ζ phase, while maintaining a constant 1 µm thickness
Γ phase.
t5: Once the δ phase reaches the surface it serves as the Zn rich side of the δ phase Fe-
Zn/steel diffusion couple, allowing for the continued growth of Γ phase at the expense of δ
phase. Cracking occurs along δ phase basal planes parallel to the coating/steel substrate
(type 2 galvanneal microstructure).8
Jordan et al.2 reported that the total Fe-Zn alloy layer growth for two different grain sizes of
an ULC steel substrate showed the same behavior over the reaction times studied. They
used the power-law growth equation Y = K.tn to estimate the growth rate of each Fe-Zn
intermetallic compound. By applying a logarithmic function in either side of the equation, it
comes to log Y = log K + n.log (t), a plot of the log value of the total alloy layer thickness (Y)
as a function of the log value of immersion time in the bath (t) can be fitted to the data to
determine a line whose slope is defined as ‘n’, the growth rate time constant value. The
growth rate time constant value is an indication of the type of kinetics controlling the growth.
An n value of 0.5 is indicative of parabolic diffusion-controlled growth, while an n value of
1.0 is representative of linear kinetics in which growth is interface controlled.
For a 0.00 wt% Al-Zn bath, the data indicated a t1/3 relationship, indicative of grain boundary
diffusion of diffusion along grain boundaries during coarsening of the grain growth. The Γ
layer growth followed a t1/4 relationship, indicative of grain boundary growth kinetics of
diffusion along grain boundaries of growing grains. The n values determined for Γ layer lie
within the range of 0.10 to 0.50. Regarding to the δ layer growth, it was reported that the n
range lies in between 0.49 to 0.65. Concerning to the ζ phase, the n range was reported to
be in between 0.16 and 0.36. Limited Zn transport along the columnar boundaries of the ζ
phase may be responsible for the observed t1/3 kinetics.
It was observed that the ζ phase formed at the steel/coating interface first, followed in time
by the formation of the δ phase between the ζ phase and the steel/coating interface and,
last, a Γ phase layer formed between δ phase and the steel/coating interface.
36
Regarding to ζ layer growth, it is most likely not significantly affected by substrate grain size
like the Γ and δ phase layers, which are more closely associated with the substrate
steel/coating interface. On the other hand, the ζ phase is affected by the rapid growth of the
adjacent δ phase layer, and its apparent consumption of the ζ phase layer also affects the
growth-time relationship determined for the ζ phase.2
Jordan et al. added that the morphology of the liquid Zn/ζ interface was non-planar which
could indicate that the kinetics of this growth reaction are governed more by the
solidification transformation from liquid to solid, as opposed to solid state diffusion
kinetics.17
On the other hand, for a 0.20 wt% Al-Zn bath, the presence of Al in the Zn bath causes the
formation of a Fe2Al5(Zn) inhibition layer. According to Ghuman and Goldstein apud Jordan
et al.2, Fe-Zn phases are able to grow in Al-containing baths, while at the same time,
Fe2Al5(Zn) is stable. δ phase formed first at the steel/coating interface followed by the
formation of Γ at the steel/coating interface. The δ phase boundary furthest away from the
steel/coating interface was in contact with the liquid Zn during immersion in the bath, and
once solidified, the liquid Zn formed solid η phase.2
Horstmann apud Marder et al.8 reported that there is an overall inward movement of the
gamma (Γ + Γ1) phase layer towards the Fe, whereas the ζ phase layer is displaced
towards the Zn melt. The δ phase layer expands in both directions, but generally towards
the Zn melt. Thus, as the Γ + Γ1 phase layer grows into the Fe, it is also consumed by the
growing δ phase layer. Similarly, the δ phase layer expands into the growing ζ phase layer
that is advancing into the Zn melt. All of these transformations are governed by the diffusion
of Zn into the Fe substrate.
Tang apud Marder et al.8 suggested that the optimum Al composition for galvannealing is
the transition from δ (or Γ) phase to the Fe2Al5Znx (η) becomes an equilibrium compound in
the bath. Al contents marginally higher than this level mean that the formation of the Fe2Al5
phase at the interface is practically impossible because of severe competition from the ζ
and δ phases.8
37
From experiments carried out by McDevitt et al.13, it was stated that the ζ phase formed in
the bath and disappeared early in the galvannealing at 500 0C. During galvannealing, it was
observed that the majority of the coating solidifies by δ formation from the liquid. The Γ1
phase formed in the bath and grew during galvannealing until all the molten Zn solidified.
The Γ phase nucleated upon dissolution of the inhibition layer as Γ1 came into direct contact
with the steel substrate.
The ζ phase generally grew in colonies of grains sharing a preferred orientation on the Ti-
Nb IF and Ti IF substrates. The number of ζ crystals on the surface of the coatings on all
substrates decreased substantially after the completion of 3s dipping. Concurrence with the
disappearance of ζ, the presence of a thin, discontinuous layer of the δ phase was
observed.
As the δ phase grew by consuming the Fe-supersaturated liquid Zn, the microstructure of
the layer was a discontinuous mixture of δ crystals and surrounding Zn liquid – in these
observations the Zn liquid has solidified upon quenching as the η phase. It appeared that
the δ phase propagated by the continued formation of small grains from the liquid directly in
front of the advancing solidification front. Even when δ grains had formed on the surface of
the coating, liquid Zn remained trapped between grains within the δ layer. Continued
galvannealing after this point resulted in the transformation of these regions of entrapped
liquid into the δ phase thus resulting in a fully compacted δ layer after the 45 s galvanneal
treatment. On the Ti-Nb IF and Ti IF substrates, this δ layer grew in thickness with the cube-
root of galvannealing time.
After the initial period of non-uniform growth, the solidification front evened out and δ growth
progressed invariably on all three substrates. However, as grains of the δ phase formed
and grew, channels of molten Zn remained among the solidified grains. These liquid
channels probably acted as avenues for fast mass transport of Fe to the advancing
solidification front. The δ solidification front progressed with galvanneal time at a rate of
thickness growth proportional to t1/3. This rate is slower than either interface controlled (∞ t)
or diffusion controlled growth (∞ t1/2).
38
They finally concluded stating that solidification of the coating at 500 0C progresses by
direct formation of δ from the liquid, instead of an intermediate step of ζ formation followed
by transformation of ζ into the δ as is often observed in solid state Fe-Zn diffusion couples.
The higher Fe content Γ phase is the only Fe-Zn intermetallic compound that can exist in
equilibrium with Fe and is predicted to form between Fe and Γ1, but its formation must be
controlled by nucleation kinetics. Moreover, a Γ1/Fe interface is a thermodynamically
unstable microstructure at 500 0C.13
2.4.2.2. EFFECT OF OPERATIONAL PARAMETERS ON Fe-Zn INTERMETALLIC
COMPOUNDS
Several production parameters have a strong influence on the Fe-Zn reactions during the
galvannealing process, as follows:
• Substrate chemistry;
• Steel pre-treatments (such as annealing);
• Zn bath composition;
• Bath temperature;
• Immersion time;
• Coating annealing temperature;
• Annealing time.10
Al is added to the Zn bath to improve corrosion resistance by either allowing for the
formation of a pure Zn overlay for galvanic protection by inhibiting the formation of Fe-Zn
phases or by introducing multiphase microstructures in the overlay coating.
However, Al control in the Zn bath is complicated by the fact that Al exists in two forms in
the bath. Some Al is dissolved in the liquid Zn phase and the rest of Al is present in the
intermetallic particles entrapped in the bath. It is the Al in liquid solution, commonly referred
to as ‘active’ or ‘effective’ Al, which can perform the function of inhibiting the Fe-Zn reaction
during galvanizing. Effective Al is also strongly dependent on the amount of dissolved Fe in
39
the bath, since supersaturated Fe can combine with Zn and Al to form Al-containing ζ, δ
and Fe2Al5Znx (η), further reducing the amount of Al in the liquid Zn.8
Baril et al. added that the Al addition in the galvanizing bath significantly reduces the rate of
the Fe-Zn reactions during both the bath immersion and the galvannealing process. This
phenomenon is called the inhibition and offers a superior way to continuously galvanize
steel sheets and produce a single η-Zn phase coating. Furthermore, this improves control
over the final alloyed coating during galvannealing.10
Perrot et al. stated that Al has been added to the galvanizing baths for many years. It
improves the lustre of the coating, reduces the oxidation rate of the Zn bath and produces
ductile layers by suppressing the growth of brittle Fe-Zn phases. However, the Fe-Zn
interactions in presence of Al becomes unpredictable due to the highly exothermic nature of
intermetallic compounds such as Fe2Al5 and FeAl3, which occur as transient phases during
the galvanizing process.24
Therefore, Al added to the bath is effective in inhibiting the formation of ζ phase in the bath
and during the heating period and also in suppressing the growth of Γ phase. Morimoto et
al. also confirmed that the volume fraction of ζ crystals decreased with increasing bath Al
content.14
Jordan et al.17, matching the results of most part of researchers, confirmed that 0.20 wt%
Al-Zn baths delay the formation of Fe-Zn phases for immersion times greater than 300 s.
Indeed, 0.20 wt% Al additions to the bath caused a non-planar growth condition of the δ/ζ
phase boundary.
Moreover, they compared coatings that were produced with different Al contents in Zn bath.
It was evident that Al in the bath reduces ‘n’ for Γ and δ layer growth.17
Pelayo et al. apud McDevit et al.15 stated that a higher strip entry temperature can have a
significant effect on the initial bath reactions because both chemical reaction rate and
diffusivity depend exponentially on temperature.
40
Regarding to dipping time, McDevit et al. reported that there was no significant difference
between morphologies of the Fe-Zn compound layer on samples dipped for 1s versus those
dipped for 3s.15
Baril et al. found that Fe content in the coating increases with decreased Al content in the
bath or with increased immersion time. For low Al bath contents (<0.12%), the effect of
immersion time is more important. In such low Al content baths, it is largely reported that
the growth of Fe-Zn intermetallic compounds is rapid, even during galvanizing. This was
confirmed by the observation of large intermetallic crystals in metallographic cross sections
of the coatings.10
According to Kanamura and Nakayama, the heating rate during galvannealing should be as
high as possible to limit the growth of both ζ and Γ phases.8
Shindo et al. found that, by increasing the galvannealing temperature and suppressing the
soaking time, the growth rate of the Γ-δ1 mixture layer was increased. This leads to a
rapidly increase in powdering. Ti and Nb, among other elements, are known for triggering
off this behaviour.25
Meshii et al. observed the influence of galvannealing time on mechanical properties. While
solidification progressed there was little increase in Fe concentration of the δ phase behind
the solidification front. The use of this behavior could be important in production where
coating mechanical properties can be optimized by producing a low Fe content δ coating.
By controlling the galvannealing time, a δ coating with a low Fe content could be
achieved.13
Dionne et al. studied the effect of cooling rate upon Fe-Zn intermetallic phases. They
observed that the cooling rate after galvanizing could also affect phase formation. A slow
cooling rate would provide time for diffusion of Zn into the phases already precipitated onto
the strip, which could account for the Fe-Al-Zn ternary phase on the Ti-stabilized IF steel.
On the other hand, fast cooling rates might promote the solidification of metastable phases
such as the Al-rich δ crystals observed on the Ti-Nb-stabilized IF steel.16
41
Bhattacharya et al., unlike many authors, stressed the importance of carrying out
experiments on work temperatures, instead of room ones. They explained that, since
segregation to surface of surface-active elements is temperature dependent, it is important
to perform analysis on samples quenched from the temperature at which steel enters the
Zn pot and not on samples at room temperature. Therefore, only segregation at the real
temperature of the steel during processing would yield true results.1
Finally, studying the influence of operational parameters on Fe-Zn-Al ternary system at 450 0C, Perrot et al. concluded that, generally, the stable Fe-Zn-Al diagrams obtained after long
reaction time are characterized by less extended single-phase regions. For instance, the Zn
solubility in Fe2Al5 and FeAl3 decreases with increasing reaction time. As the reaction time
proceeds, the system approaches an equilibrium state: the solubility of Zn in FeAl3 and
Fe2Al5, that of Al in δ and ζ, and that of Fe in the liquid are decreasing whereas that of Zn in
ferrite is increasing. On the other hand, it is not certain that equilibria between solid phases
are well established, even after 1000 h because of the low temperature (450 0C).24
2.4.2.3. EFFECT OF CHEMISTRY ON SUBSTRATE REACTIVITY AND Fe-Zn
INTERMETALLIC COMPOUNDS
The literature has shown that substrate grain boundary cleanliness is the dominant steel
substrate structural feature that controls the kinetics of Fe-Zn alloy phase formation in Al
containing Zn baths.8
The control of interstitial species begins with steelmaking and casting techniques, which are
primarily designed to minimize the amount of N and C in the steel. Subsequently, the
stabilization of N and C via precipitation reactions will be determined by thermomechanical
processing on the hot strip mill and during annealing.3
C was clearly found to retard growth rates of the δ and ζ phase layers, and have little effect
on Γ layer growth.17 Additionally, the presence of C in the steel substrate retarded the
alloying between Fe and Zn, while P in the steel favored the formation of the Γ1 phase over
the Γ phase by its surface segregation in the steel substrate.
42
Lin et al.9 have carried out TEM analysis that has clearly shown that alloying is more
progressed in the coating on IF steel than that on ELC steel, which is galvannealed at a
temperature about 5.6 0C higher than that used for galvannealing IF steel, indicating that
the C in the steel substrate retards the alloying reaction between Fe and Zn. This trend of
the observation is consistent with the early reports, indicating that solute C tends to
segregate to ferrite grain boundaries during cooling after heat treatment and restrains the
interdiffusion of Zn and Fe through ferrite grain boundaries.
They also mentioned that it was speculated that P has the same effect as C in retarding the
alloying between Fe and Zn in spite of the facts that they have the opposite effects on
intergranular fracture and that one is interstitial and the other is substitutional.9
Maschek et al. confirmed this direction. They observed that, since the onset temperatures
of some galvannealing reactions are below the melting temperature of the Zn (419 0C), the
observed intermetallic Zn-Fe growth at the interfaces has to be a solid state reaction.
Therefore, it is assumed that the retarding effect of P is based on the difficulty for P to
diffuse through solids at low temperatures.26 Marder et al. added that P segregates to the
grain boundary and slows the formation of Fe-Zn phases.8
Lin et al.9 observed that the presence of P influences the nature of the inhibition layer and
its longevity. In the coating on P-added IF steel, it is an Fe-Al compound (instead of an Fe-
Al-Zn compound on IF and ELC steels) that forms, and tends to show enhanced stability.
Furthermore, it is suggested that strong-surface, instead of grain-boundary, segregation of
P in ferrite has an important effect on the phase evolution of coating during galvannealing.9
Indeed, Maschek et al. mentioned that it is shared among most part of the authors that P
inhibits the Zn diffusion in the grain boundaries of the steel substrate and therefore prevents
the embrittlement of the steel substrate.26
Zhong et al. confirmed the work of the authors above, stating that the P solute segregated
to the ferrite grain boundaries or to the steel surface can retard the Zn diffusion and reduce
the Zn embrittlement.27
43
Allegra et al. found P to segregate to ferrite grain boundaries in LC steels (0.02 to 0.06 wt%
C) containing at least 0.04 wt% P, blocking the diffusion of Zn along the grain boundary and
lowering the thermodynamic activity. Mercer apud Allegra et al.23 proposed that the P
enrichment at the grain boundaries impeded Fe and Zn interdiffusion, thus lowering the
amount of Fe in Zn galvanizing coatings on re-P steel. Therefore, P in the base steel acts
as an inhibitor to Fe-Zn alloy growth primarily due to grain boundary segregation of P.23
Regarding to the influence of chemistry upon Fe-Zn intermetallic compounds, Shindo et al.
reported that some elements such as Mn, Se, Cu, Ca and Ti are the ones that affect the
growth of the δ1 crystal when they precipitate as sulfides on the surface. Additionally, Ti has
also been increased to predict its effect on the coating. Other elements like Nb and B affect
the crystal shape of Zn-Fe alloy and are also reported to reduce the Γ crystal growth. Their
results suggested that alloy elements added to the steel affected the Γ-δ1 mixture layer,
which developed between the substrate steel and δ1 alloy layer. 25
In order to enhance the reaction between molten Zn and the steel substrate, the addition of
Ni to liquid Zn has been considered in the practices. The addition of Ni to the molten Zn has
been made in batch galvanizing line to retard the growth of the ζ phase. This process has
become known industrially as Technigalva.20
Regarding to the most used microscopy techniques to measure segregation, the X-ray
photoelectron spectroscopy (XPS) technique is also proving to be an efficient and
convincing tool for studying the segregation phenomena that take place on steel surfaces
during annealing processes. A notable advantage of this technique is that it provides
information on the chemical state of the elements present, though it presents the limitation
of its low lateral resolution.
However, Feliu et al., performing experiments through XPS analysis, reported problems
regarding to sample contamination when in contact with the laboratory atmosphere. They
reported that, at room temperature, the surface of any metal, in contact with the
atmosphere, irrespective of its composition, instantaneously becomes coated with a thin
film of C-C/C-H groups (with a thickness less than 3 nm).4
44
In order to obtain more analytic information upon surface enrichment, GDOES-depth-
profiles have often been made. Due to the high erosion rate of this method, caused by ion
bombardment of the sample surface, it is possible to have a good idea about the chemical
distribution from the top layer to the bulk within few seconds. In this case the disadvantage
of low lateral resolution (analyzed area: 4mm in diameter) is more than compensated by an
extreme high sensitivity of the system. Therefore, GDOES is useful to detect surface
enrichments under different annealing parameters (also B and P) and not to differ between
the chemistry of various particles.28
2.4.2.4. EFFECT OF IF STEEL ALLOYING ELEMENTS ON Fe-Zn INTERMETALLIC
COMPOUNDS
Interstitial free (IF) steels have been found to have a more reactive behavior relative to
other drawing steel alloys during Zn coating processing.17 Ti added steels generally
produce a clean substrate grain boundary for faster reactions. 8
Lin et al. commented that is well known that C, Si and P in steel restrain the alloying
reaction between Fe and Zn, whereas Ti and Nb promote the reaction between Fe and Zn.
A popular explanation of the mechanism to explain the retarding effect by the alloy
elements is that they segregate to grain boundaries and limit the short-circuit diffusion of Zn
atoms into ferrite.9
Likewise, McDevitt et al. reported that P and C additions in the substrate are thought to
reduce the rate of microstructural evolution of galvanneal coatings, while the addition of Ti
is suspected to destabilize or increase the rate of development of the coating
microstructure.15
Hisamatsu postulated that the chemical nature of IF steel grain boundaries increases their
thermodynamic activity during hot-dip galvanizing, resulting in the preferential nucleation
and growth of Fe-Zn phases at these sites. Further, he suggested that a finer grain size
substrate is more reactive, because Fe-Zn phases are thought to first nucleate and grow at
substrate steel grain boundaries. Therefore, as more grain boundary area is available for
45
reaction with the liquid Zn bath on a fine grain size steel surface, more rapid Fe-Zn phase
growth results. Additionally, Jordan et al. reported that recent investigations had shown that
as the grain size of a Ti-stabilized IF steel decreases, Fe-Zn reaction rates increased during
post-dip annealing. 2
Miyasaka et al.22 observed that an increase in the amounts of Ti and Nb decreased the
incubation time significantly, while on the other hand an increase in the amounts of C and P
increased the incubation time. Since the amounts and crystallographic structure of the Fe-
Al-Zn barrier layer were not affected by the above elements, the effects of the alloying
elements on the galvannealing behavior were attributed to their influence on Fe diffusion at
the ferrite grain boundary.
Therefore, the addition of Ti or Nb shortened the incubation time for galvannealing
remarkably, though the change in the incubation time by the unit amount of Ti or Nb was
greater with Ti addition than with Nb. The decrease in the incubation time by an increase of
0.01 wt% Ti was about 1.1s, that by an increase of 0.01 wt% Nb was about 0.5s. However,
if we take into account that the atomic mass of Nb is 1.9 times as much as that of Ti, these
results indicated that the Nb showed as much influence as Ti did when their contents were
expressed in atomic percentage.22
According to Hisamatsu, Ti added sheet steel will alloy at 50-70 0C lower than a LC steel,
both being galvanized in a 0.13% Al bath at 470 0C. Mercer found that Ti-stabilized steels
were more reactive than Al killed steels for two different galvannealing temperatures. Maki
showed that the Γ phase thickness was greater than Al killed steel for every annealing
temperature studied, using a 0.11 wt% Al bath. Moreover, Ti steel inhibited the formation of
a type 1 microstructure using a 0.10 wt% Al galvanizing bath, but accelerated the formation
of type 1 with a 0.15 wt% Al bath, especially at higher galvannealing temperatures.
Similarly, Ti-Nb IF steel galvanized at 0.14 wt% Al bath accelerated alloy formation.8
Seeking some differences between the Ti and the Nb behavior upon Fe-Zn intermetallic
compounds, Pelayo et al.15 stressed that the ζ crystals on the Ti IF and the P-added
substrates grew without a noticeable preferred orientation, while on the Ti-Nb IF substrate
46
the ζ crystals grew in colonies of approximately 20-30 µm2 where all the crystals within the
colony had the same orientation.
On the one hand, the size of the ζ colonies on the Ti-Nb IF substrate were approximately
the same as the grain size of the substrate, suggesting that the ζ phase growth is
influenced by the orientation of the grains in the substrate. This implies intimate contact
between the ζ phase and the substrate, or that all phases between the substrate and the ζ
phase have an orientation relationship with the substrate grains. On the other hand, on the
Ti IF steel substrate, the surface coverage with ζ was quite variable with an equal number
of large regions (several hundred µm2) of approximately 80% coverage and similarly sized
regions with 40% coverage.15
Osman et al. reported that the initial formation of the ζ-phase (closer to the coating) for the
Ti-stabilized steels occurred without a preferred orientation, while for Ti-Nb-stabilized steels
grew in one direction and had a ‘colony’ size that was similar to the substrate grain size. It
was found that nucleation of the ζ-phase in Ti-Nb-stabilized steels preferentially occurred
over (111) α-Fe, resulting in a lower consumption of Fe and a correspondingly slower
growth rate for the Fe-rich δ- (intermediate phase) and Γ-phases. After initiation of alloying,
the presence of solute Nb at the grain boundaries may also serve to decrease the
intergranular diffusion of Fe, thus assisting in the production of a more uniform alloy
structure with a reduced Γ-layer thickness.3
Meshii et al.13 observed that, once the Fe-Al inhibition dissolved and most of the ζ phase
disappeared, the second stage of δ formation, growth of the δ phase by consumption of the
liquid Zn, could begin. Initially, growth of the δ layer was non-uniform on the Ti-Nb IF and Ti
IF substrates, but occurred more uniformly on the P-added substrate.
On the Ti-Nb IF and Ti IF substrates, propagation of the δ phase began immediately upon
reaching 500 0C because the inhibition layer had already dissolved by this point in the
processing. The start of δ propagation occurred sooner on the Ti-Nb IF and Ti IF substrates
than on the P-added substrate, but the substrate chemistry had little effect on the growth
rate of the δ phase or the Fe enrichment rate after the inhibition layer had dissolved.
47
Initially, growth of the δ phase on the Ti-Nb IF and Ti IF substrates was nonuniform.
However, this nonuniform growth of the δ phase was not characteristic of an outburst
structure.
Γ1 grew during galvannealing, and on the Ti IF and Ti-Nb IF substrates the Γ phase
nucleated and grew while the coating was solidifying. The increase in the average Fe
content of the coating is due to all of these reasons. Growth of the Γ1 phase continued until
supply of liquid Zn was exhausted. At that point, on the Ti-Nb and Ti IF substrates, growth
of the Γ1 phase ceased and Γ grew by consuming the Γ1 phase. These results indicate that
the amount of liquid Zn remaining can play an important role in the course of the galvanneal
reactions.
The Γ phase nucleated between the Γ1 phase and the steel substrate shortly after the point
in processing when the inhibition layer dissolved and the steel and the Γ1 phase interfaced
directly. On the Ti-Nb IF and Ti IF substrates Γ formation occurred within 3s at 500 0C, and
even a few Γ grains were observed after 0s at 500 0C. On the P-added substrates, Γ phase
formation was suppressed even after the inhibition layer disappeared suggesting that P
affects the nucleation of the Γ phase at the Γ1/Fe interface.13
Marder et al. added that Ti and Ti-Nb solute additions, which enhance grain boundary
reactivity, resulted in more rapid growth kinetics of the Γ and δ phases than an ULC steel
for a series of IF steels in Zn coatings containing an Fe2Al5 inhibition layer (Zn-0.20 wt% Al
bath).8
2.4.2.5. EFFECT OF SURFACE OXIDES ON Fe-Zn INTERMETALLIC COMPOUNDS
Mn is the alloying element that is found in the greatest proportion on the annealing steel
surface. According to the literature, it seems rather usual to find a continuous and thin film
of Fe oxide over the steel surface, while Mn and Si oxides may be distributed in the form of
an island.
The strong surface enrichment in Mn could be explained in general as due to equilibrium
segregation and/or external selective oxidation. Selective oxidation of reactive alloying
48
elements appears to be a common phenomenon in the steel annealing treatments under
vapor-containing protective atmospheres.
The increase in Fe0 should be caused by the greater affinity of Mn for O, which could
reduce the Fe oxides in the thin layer present on the outer surface of the steel. In
accordance with the idea of Mn oxidation by the Fe oxides formed in the annealing process,
Vanden Eynde et al., working in a nonreducing atmosphere (dry N), found that the
reduction of these oxides did not come from the reactions between the annealing
atmosphere and the steel surface but from the interchange of O between the Fe oxides and
the alloying elements that segregate on the surface (Mn and Si).
Some authors apud Feliu et al.4 have found a dear relationship between the inhibition of the
growth of Fe-Zn intermetallic compounds and the fraction of the steel substrate surface
covered by Mn oxides as a result of the annealing process.
The literature mentions a tendency toward the formation of outer layers of MnO during the
annealing process of IF steels even in a reducing atmosphere. Feliu et al. reported that, on
annealed IF steels (0.6 wt% Mn), the element that exhibited the greatest tendency towards
segregation was Mn, which reached contents on the surface of close to 25%. The lower C
content in the IF steel compared with the LC steel may increase the tendency for Mn to
diffuse during the annealing process, in concordance with the observations of Zhang and
Bensinger.4
Moreover, it is well known that the metallic alloying elements segregate toward the surface
of IF steels in order to precipitate in the form of metallic oxides and hydroxides as a result of
the annealing process. This segregation seems to inhibit the nucleation and growth of Fe-
Zn intermetallic phases.29
Additionally, the Si contents in the bulk steels are lower than the contents of other alloying
elements such as P, Cr, Al, Ni, etc., which are not detected by XPS on the outer surface of
the annealed steels. This fact agrees with the great tendency of Si to precipitate on the
surface of steels in the form of fayallite (Fe2SiO4) during some annealing cycles.4
49
Perez-Ravenga et al.29 observed and studied the surface contamination of steel, which has
endured the annealing operation and persists even after galvannealing treatment. It is as if
the contamination had shifted during this treatment from the surface of the annealed steel
to the outer surface of the galvanneal coating (Fe-Zn alloy phase), which had grown on that
surface.
In the case of the galvanneal coating, the growth of the Fe-Zn alloy would push the
contamination toward the outer surface of the coating, while on the galvanized coating it
would remain at the Fe-Zn alloy/Zn interface, and thus be impossible to detect. This
phenomenon is somewhat similar to other cases reported in the literature, e.g. that in Zn
(Al) baths, the Fe2Al5 inhibiting layer is pushed into the melt during the formation of the
galvanized coating.29
2.4.3. CRATER FORMATION
Marder et al.8 pointed out that craters were found to be inherent to the galvanneal coating
layer, causing major fluctuations in coating thickness and appear to be associated with
outburst formation. They proposed the following mechanism to explain the relationship
between outburst formation and the appearance of craters:
After an outburst of Fe-Zn intermetallic compounds occurs at the substrate grain boundary,
rapid growth of these solid phases occurs into the liquid. The liquid Zn will remain present
between the crystals, which enables very fast transport of Fe from the substrate. Capillary
effects of the pores and channels between the crystals drain the nearby regions where no
alloy reaction takes place, whereby ‘dry spots’ or craters are formed. Indeed, more craters
were seen on the surface of galvanneal coatings for low Al baths, which were related to
outburst formation.8
Faderl et al. reported that cratering occurs preferentially at low galvannealing temperatures
as well as low ‘galvannealing-reactivity’ (re-P IF steel), and therefore indicates good
adhesion of powdering performance. This is due to the formation mechanism of craters
(which was explained by them as microscopic Zn transfer from ‘low-reactivity’ areas to
‘high-reactivity’ ones). Moreover, in several cases different coating formation (cratering)
50
was observed over the width for one coil but the powdering behavior was not influenced by
cratering.30
van der Heiden et al.12 pointed out that, though hard to recognize by SEM 2D photography,
secondary electron 3D images clearly revealed that all the galvanneal surfaces were
covered with craters – which they characterized as discontinuities. Typically, those craters
covered 5-10% of the coating surface, as was quantified by measuring the thickness
distribution by automatic image analysis. By applying SEM stereophotography, the
columnar and compact crystals of the coating layer could be observed, in particular inside
the craters on bottom and walls. No indications of fractured surface could be found inside
the craters. Furthermore, SEM photographs and EPMA showed that there was always
some Zn left on the bottom of the craters.
After became clear that the craters were not the result of sample preparation, it was
wondered if skin passing might cause crater formation. However, craters were also found in
samples from a coil that had not been skin passed. Even after 2% skin passing of this
material on a laboratory mill, the same amount of craters was found. Indeed, only the
surface was more flattened. Moreover, there were no fracture indications inside the craters.
After studying many galvanneal products of several commercial lines, it turned out that
craters are inherent to the galvanneal coating layer. It had become clear that they cause
great differences in thickness on microscale and are formed during the galvanneal process
itself.
One mechanism of crater formation was proposed by van der Heiden et al., as follows:
Before the strip is entering the galvanneal furnace, an alloy layer of Fe2Al5 has been formed
at the interface. This Fe2Al5-layer retards Fe-Zn alloying. When locally breaking up of this
layer is promoted by a temperature rise, the Fe-Zn alloy reaction at this point proceeds very
rapidly, leading to the appearance of the well known outbursts of Zn-Fe crystals. As
suggested by Rensen and van Eijensbergen on the basis of the hot-stage microscopy
experiments, liquid Zn remains present between the crystals, which enables very fast
transport of Fe. Capillary effects of the pores and channels between the crystals will drain
the nearby regions where no alloy reaction takes place, whereby ‘dry spots’ or craters are
51
formed. Obviously, at alloy/outburst locations, the interface moves inwards, as a result of
Fe consumption.12
Finally, they concluded supposing that craters might have a positive effect on mechanical
paint adhesion as well as on anti-powdering properties. The latter because powdering is the
result of compressive stresses in the coating. The craters offer the possibility to prevent the
build up of these compressive stresses to some extent. So the presence of craters has no
negative influence on the product performance of galvanneal coatings.12
van Koesveld et al.31 pointed out that the occurrence of outbursts along ferrite grain
boundaries was shown to be the driving force for deep crater formation in the galvanneal
coating. It causes, among other things, a non-uniform consumption of Fe from the substrate
interface grains during galvannealing.
Through SEM back-scattered 3D electron images they were able to show very clearly the
absence of fractured crystals on the wall and the bottom, indicating that the craters are not
created by deformation like skin passing, confirming van der Heiden’s assumption12.
Consequently, their origin must be sought for in the galvannealing process itself.
Moreover, SEM and EPMA measurements could show that some residual Zn layer is
always present at the bottom of the craters, which thus remain adequately protected
against corrosion. Additionally, they found that free Zn remains present between the
needle-like crystals of the outbursts. The latter crystals grow thicker than the surrounding
residual liquid Zn, which is sucked into the open outburst crystal structure and on top of it
by capillarity and surface tension effects. The outcome of this draining effect is the
formation of craters.
They also mentioned about an additional crater formation mechanism, which had recently
been put forward and will be explored following this by O’Dell et al. This theory is based on
epitaxial relationships prevailing between ferritic substrate and galvanneal coating.
Providing that coating Al content is high enough, Fe-Zn reactions are supposed to be slow
on (111) surface ferritic grains, while they proceed faster elsewhere. According to this
52
mechanism, craters are then preferentially formed on (111) ferritic grains, the size of which
controls crater dimensions.
Finally, they concluded also stating that craters have no negative influence on galvanneal
product user’s properties, including corrosion. On the contrary, the presence of craters has
a positive effect on properties like coating adhesion.31
Similarly, O’Dell et al.19 stated that the presence of craters within the galvanneal coating
had been reported to reduce the amount of powdering experienced by the galvanneal
coating. They reported that two different types of crater were observed, a light and a dark
one. Within the light crater ordered crystals were seen at the base of the crater. The Zn
content of these crystals is very similar to that found over the rest of the galvanneal coating.
On the other hand, the dark crater shows a large presence of Fe. They also reported that,
on all the samples analyzed by their study, no craters were found to be present after
deformation through a standard cupping test.
Regarding to the influence of operational parameters upon the occurrence of craters, they
found that the process parameters which have the most effect upon the crater coverage are
the strip entry temperature (SET) and the temperature after dipping (TAD), which is defined
as the lowest temperature the panel reaches after it has been dipped and prior to
galvannealing. It was reported that an increase in the SET leads to an increase in crater
coverage due to the improved nucleation ability of the ζ phase at these higher temperatures
and the easier formation of the ordered ζ crystals during the initial dipping period. On the
other hand, a decrease in the TAD promotes the formation of craters within the galvanneal
coating because a lower TAD promotes the growth of the ordered ζ crystals resulting in a
more compact and more stable structure.
They explained that diffusion of the Fe atoms into the ordered ζ crystals is limited by either
the orientation relationship between the craters and the (111) substrate grains due to the
persistence of an inhibition layer between the ζ crystals and the substrate. Moreover, the
Zn above the ordered ζ crystals is drawn into the surrounding developing galvanneal
coating so that the crater is created.
53
Briefly, as mentioned earlier, O’Dell et al.19 pointed out that craters within the galvanneal
coating help to reduce the powdering tendencies of the galvanneal coating by inhibiting
crack propagation along the weak Γ/substrate interface. Indeed, they alleviate the
compressive strains by the failure of the ordered ζ crystals, and therefore prevent the
coating from powdering. Summarizing, they act as stress relievers for the galvanneal
coating when subjected to compressive strains.19
2.4.4. EFFECT OF IF STEEL ALLOYING ELEMENTS ON GALVANNEALED COATING
STRUCTURE
With the increased utilization of hot-dip galvanized coatings, IF steels became commercially
viable since these grades did not require a pre- or post-anneal to prevent strain aging.
Today, hot-dip galvanneal coated IF steels processed through continuous galvanizing lines
predominantly contain a combination of Ti and Nb for interstitial free stabilization.
Additionally, some hot-dip galvanize coated sheets as well as batch annealed sheets (i.e.,
cold rolled or electrogalvanized sheets) contain Ti and Nb, especially higher tensile strength
grades. The remaining IF steels currently produced rely upon Ti additions alone for
interstitial stabilization.
Osman and Garcia have reported that, in addition to alloy phase development, Nb additions
are also beneficial for reduction of ‘woodgrain’ appearance on the surface of the galvanneal
sheet. The use of Nb for interstitial stabilization therefore reduces the necessary Ti content
in the sheet, thus reducing the propensity for this type of surface appearance.3
2.4.4.1. STEEL SUBSTRATE REACTIVITY
Interstitial Free (IF) steels have been found to have a more reactive behavior relative to
other drawing quality steel alloys during Zn coating processing. The reactive behavior of IF
steels can lead to difficulty in controlling the degree of alloying that occurs during post-dip
annealing. Rapid reaction times during hot-dip galvanizing and post-dip annealing can lead
to overalloying and subsequent poor formability properties of the coating during press
forming operations.17
54
Marder et al.8 stated that because IF steels are more likely to form carbide, nitride, sulfide
and phosphide precipitates in the grains, preventing segregation to the grain boundaries,
the concept of excess Ti is very important and was introduced through the following
formula:
excess Ti = total Ti – 3.99C – 1.49S – 3.42N – 1.55P
Where a positive excess Ti exists, it indicates a clean, carbide free and reactive grain
boundary. A negative value of excess Ti would indicate that not all of the solute C is tied up
and Zn diffusion down the boundaries would be blocked. Hence, the cleaner is the
substrate grain boundaries, the easier Zn atoms diffuse along them.8
Hisamatsu apud Jordan et al.17 confirmed that the substrate grain boundaries in IF steels
are more typically reactive because they are essentially C free due to carbide and
carbonitride formation that results from carbide stabilizing additions of Ti and or Nb. Indeed,
according to his theory, the nucleation of Fe-Zn phases first occurs where the Fe-Al
inhibition layer first breaks down, at a thermodynamically active steel substrate grain
boundary.
Similarly, Jordan et al. affirmed the high reaction rates normally associated with IF steels
may be related to their ultra low C content. They added that most of the research to date
concerning substrate and Zn bath alloying additions had been conducted on pure Fe
substrates.17
Dionne et al.16 reported that C and P in the substrate are thought to have a retarding effect
on the alloying reaction between Fe and Zn during galvannealing. It has been proposed that
these elements segregate to the grain boundaries and limit the short-circuit diffusion of Zn
atoms into the ferrite substrate. On the other hand, Ti additions promote the galvannealing
reaction, possibly by reducing the concentration of free C.
Moreover, they found that the area fraction covered by δ crystals and the extent of reaction
in the regions surrounding the emerging ferrite grain boundaries on Ti and Ti-Nb IF steels
55
partially galvanized for 4 s were similar, which indicates that the reactivity of both substrates
at the beginning of the galvannealing reaction was comparable.16
Zhong et al.27 explained that relatively slow lap shear strength for the Ti IF steel coatings
tested is caused by the excess Ti in the steel. Since excess Ti results in clean grain
boundaries by fixing all the C, S and N solute atoms within the ferrite grains, Zn diffusion
along the ferrite grain boundaries is facilitated during galvannealing, and these atoms can
lead to a substrate grain boundary embrittlement, thus reducing the bonding strength.
They added that, for Ti IF steel coatings, there are more ferrite grains being pulled-out,
compared with the Ti-Nb IF steel coatings at a given coating Fe level. This proves that Zn
embrittlement is the most severe in Ti IF steels, resulting in relatively low coating/steel
bonding strength. With respect to failure, it was reported that Ti promotes flaking failure due
to more severe substrate Zn embrittlement.27
van Koesveld et al.31 pointed out that diffusion hampering by interstitials, through
obstructing diffusion paths – dislocations and grain boundaries – with soluble C or N atoms,
might take place in the Ti-Nb IF steel. Its chemistry sometimes corresponds to a ‘negative
excess Ti’ suggesting that some limited amount of C and/or N can have been left free to
segregate to ferrite grain boundaries.
They found that the weight losses observed in the case of Ti-Nb IF steel are systematically
somewhat lower and less scattered than those of their Ti IF steel counterparts. Associated
Rhesca simulation test results strongly suggested that this difference in performance
derived from the very distinct surface reactivities of the two steel grades considered. Based
on cup tests carried out on identically treated industrial products, it appeared that Ti-Nb IF
steel grades exhibited a somewhat better and less scattered powdering resistance than Ti
IF steel grades.
Furthermore, it was reported that in the case of the Ti-Si IF steel grades, galvannealing time
had to be lengthened from 25 seconds to 40 seconds to produce the same degree of
alloying as in its Ti IF steel counterpart, because of its lesser overall reactivity. Besides, in
the re-P Ti-Nb IF steel, Zn penetration into ferrite grain boundaries is prevented by the
56
combined effects of TiFeP precipitation and interstitial atoms (excess Ti is even more
negative). However, no surface analysis performed in their study confirmed other authors’
assumption that P could segregate at the free surface and thus retard Fe-Zn reactions.31
Regarding to outburst formation on the surface of IF steel galvanneal coatings, Miyasaka et
al. mentioned that it is well known that galvanneal of IF steel is prone to outburst
microstructure and galvannealing of IF steels proceed rapidly compared with that of the
conventional Al-killed steels.22
Marder et al. also reported that Ti-IF steels are more sensitive to outburst formation than Al-
killed or even Nb-Ti steels. 8
Osman et al. confirmed that while this outburst phenomenon occurs for both types of steels,
Ti-stabilized IF steel grades exhibit a higher propensity for outbursts than Ti-Nb-stabilized
IF steel grades. As a result, the growth of Fe-rich phases is more rapid for Ti-stabilized IF
steels than for Ti-Nb-stabilized IF steels, resulting in a deleterious increase in the thickness
of the Γ/Γ1-layer (the phase closer to the substrate).3
Bhattacharya et al. mentioned that it has been suggested that since the affinity of Nb and P
is lower than that of Ti and P, Nb does not scavenge P, as does Ti through the formation of
FeTiP (Fe-Ti phosphides). As a result, some P is left reducing the outburst reaction.
Moreover, they added that another school of thought is that the outburst reaction is higher
in Ti IF steels because it has cleaner grain boundaries than Ti-Nb IF steels. This is a result
of the segregation of Nb to the grain and sub-grain boundaries in Nb containing IF steels.
Finally, one further idea is that Nb also segregates to the surface, thus affecting the
formation of the inhibition layer and, in turn, the galvannealing behaviour.1
2.4.4.2. FORMATION OF OUTBURSTS
Galvannealing reactivity is the annealing time necessary to reach 8.5 % Fe in the coating.30
It is significantly increased if the annealing is carried out below the recrystallization
temperature, for high recrystallization annealing atmosphere dew points – known to lead to
internal selective oxidation, with the formation of nodules below the surface - result in
57
increased reactivity.31 Consequently, the rapid penetration of Zn at grain boundaries leads
to the formation of outbursts, which are essentially regions of more rapid alloy growth.3
Guttmann apud Marder et al.8 proposed that the Fe2Al5 diffusion short circuits would
coincide with emerging substrate grain boundaries. Because of the Fe2Al5/substrate steel
orientation relationship, the high angle boundaries between colonies of the inhibition layer
could be coincident with random substrate grain boundaries. Oxide/Fe2Al5 interface short
circuits may also be preferentially situated at substrate steel grain boundaries. Thus, both
the effect of steel solute additions of ferrite grain texture and grain size can affect the rate of
inhibition layer breakdown of outburst formation.8
Indeed, his studies meant that the Fe2Al5 inhibition layer that developed on the substrate
during galvanizing forms epitaxially so that Fe2Al5 crystals develop colonies of the same
orientation as the underlying substrate steel grain. The colonies form low-angle grain
boundaries when contained within an underlying substrate grain interior and, therefore, Fe
and Zn diffusion follow normal diffusion rates at these low-angle boundaries. However, at
the location of a substrate grain boundary, two adjacent Fe2Al5 colonies of different
orientation form a random high-angle boundary. This high-angle boundary can become a
short circuit diffusion path for Fe and Zn interdiffusion, which leads to Fe-Zn alloy phase
outburst formation (rapid Fe-Zn phase growth) at substrate grain boundary sites. Thus, the
steel substrate grain boundary is expected to be indirectly related to outburst formation,
because the steel substrate grain boundary fixes the location of the short circuit diffusion
paths in the Fe2Al5 inhibition layer.2
According to Hisamatsu apud Marder et al.8, Fe-Zn reactivity at grain boundaries of the
substrate will depend upon the ability of the solute element to segregate to these sites.
These alloying additions can be separated into elements that segregate to the grain
boundaries (e.g. C, P), and those that will form compounds (e.g. Ti, Nb) that will precipitate
throughout the grain, leaving the grain boundary ‘pure’ or ‘clean’. Clean grain boundaries
will have no barrier to Fe-Zn compound formation, while segregated boundaries will reduce
the thermodynamic activity at these sites, decreasing outburst and Fe2Al5 inhibition layer
breakdown.8
58
Maschek et al. apud Marder et al.8 also reported that P was found to retard the formation of
outbursts and localized growth of the Fe-Zn phases.21 Marder et al. complemented stating
that P was found to segregate to ferrite grain boundaries in re-P LC steels, blocking the
diffusion of Zn along grain boundaries and lowering the thermodynamic activity. They
mentioned that it had also been proposed that P segregates to the steel surface during
recrystallization annealing, stabilizing the inhibition layer and retarding the rate of Fe-Zn
phase growth reactions during galvanizing.8
Regarding to IF steels, Maschek et al. found that they, in contrast to LC steels, tend to form
outbursts preferentially at the grain boundaries of the steel substrate. The retarding effect of
P is therefore explained by the formation of Ti-P-precipitates near the grain boundaries of
the IF steel substrate. These precipitates consume a distinct amount of solute Ti and
therefore compete with the formation of TiC in the steel substrate, allowing additional solute
C to segregate at the grain boundaries, which will subsequently retard both the onset of the
Fe-Zn reaction at the interface and the formation of outbursts. Other theories suggest that
the P diffuses to the surface of the IF steel substrate or interacts with the Fe-Al-Zn inhibition
layer after segregating to the grain boundaries in the steel substrate.26
Moreover, Marder et al. observed that outburst formation readily occurred in the 0.20 wt%
Al baths due to the initial formation of the Fe2Al5 inhibition layer; however, increased grain
size significantly retarded incubation time and promoted Fe2Al5 inhibition layer stability.8
Nevertheless, McDevitt et al.14 affirmed that there was no observation of rapid Fe-Zn growth
at the inhibition layer/substrate interface and no mechanical breaking of the inhibition layer
has been observed under these conditions simulating commercial galvanneal processing.
They said that it proves that outburst mechanism is not the sole means for inhibition layer
breakdown during galvannealing. Additionally, they pointed out that if the outbursts
mechanism is observed to occur under similar galvannealing conditions in the future, then it
will be clear that the inhibition layer is subject to breakdown by either or both of the
competing mechanisms.
In addition, the presence of Γ1 on the inhibition layer prevents intimate contact of the molten
Zn with the inhibition layer, which in turn may prevent outburst formation. On long dipping
59
time samples or on samples produced in higher Al content baths, less surface area of the
inhibition layer is covered by Fe-Zn compounds so that the inhibition layer is less protected
from direct attack by the liquid Zn, and outburst formation may occur.
Finally, they concluded affirming that their present results also demonstrated that
phenomena related to substrate grain boundaries cannot be exclusively attributed to
outburst formation as grain boundaries also played an important role in the Al diffusion into
the substrate during inhibition layer dissolution.14
Meshii et al.13 also confirmed that a correlation between formation of the Γ1 phase and
occurrence of outbursts exists. They started affirming that, in addition to the Fe-Al layer
inhibiting the interaction between the substrate and the molten Zn, the Γ1 layer may too
have hindered molten Zn-substrate interaction and further formation of the Fe-Zn
intermetallic compounds. The presence of the Γ1 layer after the dipping process may in part
explain why outbursts were not observed. Reporting to Hisamatsu, they concluded stating
that the outburst phenomenon occurs by direct reaction of the galvanizing bath and the
steel substrate through an Fe-Al-Zn ternary compound layer formed in the initial stage of
immersion. Finally, as the Fe-Al inhibition layer broke down, the Γ1 layer prevented
interaction between the substrate and the molten Zn, which in turn deterred rapid localized
growth of the Fe-Zn compounds.13
Baril et al.10 stated that outbursts are more likely to form at the grain boundaries of the
substrate. In fact, random high angle boundaries of Fe2Al5 are short circuit diffusion paths
for Zn. They further observed that the formation of outbursts is expected to lead to a
significant increase of the local Fe content in surrounding liquid Zn, which will tend to pump
Al from the bath to form Al-Fe-rich compounds. This may explain the relatively large amount
of Al in the coating for immersion time of 3.5 and 6 seconds, for an Al bath content of
0.10%.10
Through higher magnification analysis, van Koesveld et al.31 reported that the presence of
ferritic grain boundaries in the depressed areas was revealed, which correspond to initial
outbursts. This substantiates that outbursts appear most generally along the ferritic grain
60
boundaries of IF steel substrates, thus accounting for the apparent relationship between
ferritic grain and crater sizes.
Moreover, obviously enough for them, outbursts could only grow by consuming the Fe
immediately underneath them, which results in an average interface depression of about
10% of the total coating thickness (thus accounting for the total amount of Fe needed for
complete alloying). The corresponding enhanced Fe consumption along the grain
boundaries increases the steel-coating interface roughness, which improves its shear
strength.31
Regarding to the morphology of outbursts, Zuhr et al.23 observed that it was similar to that
found on other 0.20 wt% Al-Zn coatings and consisted of a thin Γ layer of approximately 1.0
µm at the interface and a thick 10- to 100-µm δ phase layer adjacent to the Γ layer. Similar
to other 0.20 wt% Al-Zn coatings, the δ layer was in contact with the liquid Zn. Furthermore,
the localized Fe-Zn phase growth was found to correspond directly to the location of a
substrate grain boundary site, and the results show that nucleation of the outbursts
occurred in the same grain boundary independent of P surface segregation.23
Jordan et al.2 added that the Fe-Zn outbursts typically had trailing elongated tails of Fe-Zn
phase that extended down along substrate grain boundaries. They also confirmed that the
Fe-Zn outbursts also appear to consist of Γ and δ phase layers.2
2.4.4.3. FORMATION OF CRATERS
It was postulated that Ti addition to the IF steel gives superior mechanical properties
compared to Nb added IF steel. Ti added IF steel produced in a very simple process
exhibits superior mechanical properties. However, Ti addition tends to cause streaky
surface defect, which appears after galvannealing as white band along rolling direction. The
appearance of streaky defect is depressed with the decrease in Ti content of substrate.
Therefore, the amount of Ti addition must be kept at lower level (about 0.025%).32
Miner et al.33 reported that streaks are generally associated with Ti-stabilized IF steel
substrates and are caused by differences in reflectivity associated with local variations in
61
coating morphology. Countermeasures for streaking include lower pot Al levels to promote
more uniform alloying, lower Ti levels to avoid nonuniform scaling during hot rolling and
better control of upstream processing to provide a more uniform surface for the
galvannealing reaction. It was also affirmed that galvannealed steels are more susceptible
to electroprimer cratering which can adversely affect pain line productivity and under some
conditions can exhibit loss of the galvanneal coating during paint chipping.33
Finally, Parks et al.11 studied the influence of ferritic rolling upon the occurrence of streaks.
They pointed out that ferritic rolling of ultra-low C interstitial-free steels has a positive effect
on Ti-streak surface defects that occur on Zn-coated products. They characterized this
defect as associated with Ti-containing surface oxides that appear as dark streaks following
galvannealing. While no quantitative data are available, it can be concluded that lower
furnace temperatures result in fewer surface oxides.11
62
3. EXPERIMENTAL PROCEDURE
3.1. MATERIALS
Two IF steel grades were used for this project, a Ti-stabilized and a dual (Ti-Nb) stabilized
IF steel grade. Table 4 features information on the steel samples used for this project.
Table 3. IF Steel substrates and conditions tested.
Steel Substrate Condition Code
Ti-stabilized IF steel As received (cold rolled) A
Dual-stabilized IF steel As received (cold rolled) B
Ti-stabilized IF steel Galvanized (GA) AG
Dual-stabilized IF steel Galvanized (GA) BG
120 x 200 mm cold rolled steel sheet coupons were received from two different suppliers,
where the longitudinal axis corresponds to the rolling direction. The as-received coupons
were immediately degreased with soap and dried off with ethanol, and a 50 x 50 mm central
surface area was grinded and polished on each sample for the purpose of minimizing the
surface roughness for further GD-OES surface analysis.
The chemistries for the two IF steel grades, obtained by the Jobin-Yvon/Horiba Quantum IQ
V2.22 Glow-Discharge Optical Emission GD-OES in bulk composition mode, are shown in
table 5.
63
Table 4. Chemistries for the Ti-stabilized IF steel grade and the dual-stabilized IF steel grade.
Ti-stabilized IF
Steel
Dual-stabilized
IF Steel
Ti-stabilized IF
Steel
Dual-stabilized
IF Steel Steel
Chemistries Steel A Steel B Steel AG Steel BG
wt% C 0,003 0,002 0,002 0,002
wt% Mn 0,16 0,19 0,11 0,18
wt% P 0,008 0,005 0,015 0,009
wt% S 0,008 0,005 0,005 0,006
wt% Si 0,007 0,005 0,001 0,009
wt% Cu 0,020 0,018 0,008 0,020
wt% Ni 0,013 0,019 0,004 0,010
wt% Cr 0,02 0,02 0,01 0,05
wt% Mo 0,004 0,001 0,002 0,007
wt% V 0,002 0,001 0,010 0,001
wt% Nb 0,006 0,029 0,005 0,020
wt% Ti 0,052 0,020 0,043 0,030
wt% Sn 0,002 0,006 0,001 0,002
wt% Al (sol.) 0,03 0,035 0,030 0,060
wt% N 0,002 0,003 0,002 0,001
exc Ti* (%) 0.009 (0.013) (0.002) (0.004)
* exc Ti = Ti – 3.99.C – 1.49.S – 3.42.N – 1.55.P
64
3.2. TESTS PERFORMED
Table 6 features the tests the above mentioned steel samples were submitted to.
Table 5. Tests performed on the IF steels.
Steel Code
Characterization Tool A B AG BG
Sample Chemistry X X X X
HDG Simulations X X
Interrupted Cycle HDG Simulations X X
Glow Discharge Optical Emission Spectroscopy X X X X
Scanning Electron Microscopy X X X X
X-ray Photoelectron Microscopy X
Figure 1 features a chart displaying a comprehensive plot of the steps carried out by this
research project including sample characterization as well as microscopy techniques used.
Figure 1. Flowchart of the steps carried out by this research project.
S t e e l A S t e e l B
I F S t e e l B a r e S a m p l e s
A s - r e c e iv e d S a m p l eC h a r a c t e r iz a t io n
X P S A n a l y s i s
S E M / E D S A n a l y s i s
H D G S i m u l a t io n sw i t h G A In te r r u p t e d C y c l e s
G D - O E S A n a l y s is
H D G S i m u l a t io n s
S t e e l C S t e e l D
I F S te e l G A S t r e a k e d S a m p l e s
A s - r e c e i v e d S a m p l eC h a r a c t e r i z a t i o n
G D - O E S A n a l y s is
S E M / E D S A n a l y s is
O p t i c a l M ic r o s c o p eA n a l y s i s
AG BG
65
3.3. METALLOGRAPHIC CHARACTERIZATION OF AS-RECEIVED
SAMPLES.
IF steel sheet coupons were sheared down to 10 x 10 mm-size samples and then mounted
as three-piece cross section sandwiches in cold resin for the purpose of metallographic
characterization. As soon as the resins solidified under vacuum environment, the samples
were ground on Si carbide papers starting from 240- down to 600-grit papers in water. After
that, the samples were polished on rounded Si carbide polishers starting from 800 down to
4000-grit papers, and then properly dried off to be taken to next sample preparation stage.
The final fine polishing step was carried out by using both 3µ- and 1µ-grit diluted Al2O3
paste .
Etching was performed by immersing samples for 1 min in a 1% HNO3-Ethanol (nital)
solution. Next, the samples were taken to an optical microscope for their first
characterization. For the purpose of measuring the average ferrite grain size, an imaging
software attached to the Carl-Zeiss optical microscope was used. The methodology
complied with the guidelines according to the Heyn Lineal Intercept Procedure described on
the ASTM E 112 - 96 standard.
3.4. IF STEEL GA SAMPLES WITH STREAKS
Industrial IF steel GA samples with streaks were received from two different suppliers and
correspond to both Ti- and dual-stabilized IF steel grades. GD-OES bulk composition mode
was used to obtain the bulk steel chemistry of the samples.
Knowing the chemistries for the two sets of industrial streaked samples, next step carried
out was the coating characterization, when GD-OES surface analysis method was used for
that purpose. The author sought to establish comparisons between dark and light streaked
areas for both Ti- and dual-stabilized IF steel GA samples.
66
Representative dark and light streaked areas on industrial samples supplied by two
different suppliers were chosen for morphology characterization and chemistry analysis.
The samples were degreased with soap and cleansed with ethanol, sheared to fit the GD-
OES sputtering beam size and then stored in desiccators to keep them from excessive
humidity.
3.5. HOT-DIP PROCESS SIMULATOR
GENERAL DESCRIPTION
In order to simulate the operating conditions of a commercial Continuous Galvanizing Line
in a smaller scale, an Iwatani-Europe Hot-Dip Process Simulator, HDP, was used. This
equipment is composed of an upper chamber for sample input/output and further cooling by
faster rates. The infrared heat treatment furnace lies in the middle of the equipment,
operating by either radiant or induction heating. Finally, the lower middle part of the
equipment consists of a chamber for slower cooling rates, which is placed right above the
30 kg Zn-bath crucible.
The picture below displays the Iwatani-Europe Hot-Dip Process Simulator used to perform
HDG simulations upon IF steel panels.
67
SAMPLE GEOMETRY
In order to fit the HDP simulator sample input chamber, the cold rolled IF steel sheets were
cut off in several coupons 120 x 200 mm in area. Next, a 50 x 50 mm central area was
polished in order to minimize the surface roughness of the samples.
SAMPLE PREPARATION FOR THE HDP SIMULATOR.
Industrial as-received cold rolled Ti-stabilized and dual-stabilized IF steel samples were
received from steelmakers.
Due to dimensional restrictions given by the sample input chamber at the HDP simulator,
the samples were sheared down to 120 x 200 mm coupons.
Furthermore, in order to prepare the sample surface for further HDP simulation, soap was
used to remove grease as well as ethanol for contamination removal.
Moreover, tap water was used to wash residues away from the sample surface, followed by
ethanol application on the sample surfaces in order to remove any contaminants.
Finally, fast hot-air cooling was performed in order to keep the sample surfaces from
excessive ethanol residues.
3.5.1. DEFINITION OF HEAT TREATMENT CYCLES FOR Ti- AND DUAL-STABILIZED
IF STEEL GRADES.
In order to set up the heat treatment cycles for both IF steel grades with a special focus on
the galvanneal heat treatment parameters - GA temperature and soaking time -, trial
dipping samples were run at the HDP simulator aiming to produce a fully-developed GA
coating on top of the steel substrate.
3.5.1.1. EXPERIMENTS FOR SETTING GA TEMPERATURES AND SOAKING TIMES
68
The experiments were carried out by using the same effective Al 0.13 wt% in the Zn pot
and same furnace controlling atmosphere 95% N2 + 5% H2 in the cooling chambers. The
operating parameters were set up to produce a 1 µm-thick Γ layer close to the interface
between the Zn coating and the steel substrate.
3.5.2. INTERRUPTED CYCLES FOR IF STEEL BARE SAMPLES
Once the thermal profiles for both IF steel grades were properly set up, interrupted cycles
were run for bare samples from both IF steel grades, stopping the cycle at pre-determined
spots along the thermal curve, and samples were taken out at each one of these spots
along the thermal profile. The interrupted thermal cycles for both IF steel grades can be
seen on graphics of Figure 3, whose main processing parameters are as follows:
Thermal cycle set up for samples of both IF steel grades are given according to table 7.
Table 6. HDG operating parameters carried out for setting up thermal cycles run on both IF steel grades during HDG simulations.
Operating parameter Ti-stabilized IF steel grade Dual-stabilized IF steel grade
Peak Annealing
Temperature (PAT) 875 0C 830 0C
PAT Soaking Time 22,1 s 61,2 s
Zn Bath Temperature 465 0C 465 0C
Zn Bath Immersion Time 5,1 s 5,8 s
Galvannealing Furnace
Temperature 510 0C 510 0C
Figure 2 shows plots of the thermal profiles employed.
69
Heat Treatment Cycle (Ti-stabilized IF steel samples)
0
100
200
300
400
500
600
700
800
900
0 50 100 150 200 250 300 350 400Time (s)
Tem
per
atu
re (
deg
C)
Heat Treatment Cycle (dual-stabilized IF steel samples)
0100200300400500600700800900
0 50 100 150 200 250 300 350Time (s)
Tem
per
atu
re (
deg
C)
Figure 2. Thermal profiles set up for both IF steel grades at the HDP simulator. The markers show sampling conditions.
The purpose of this procedure was to observe how Ti and Nb surface concentrations evolve
along the thermal profile for both Ti- and dual-stabilized IF steel grades compared with the
as-received condition as well as to assess the effect of the annealing and galvannealing
heat treatments upon the Ti and Nb surface segregation.
3.5.3. Ti-STABILIZED IF STEEL GRADE
12 (twelve) Ti-stabilized IF steel coupons were produced throughout the thermal profile set
up for Ti-stabilized IF steels by interrupted cycles at the HDP simulator, one from each spot
where the thermal cycle was purposely stopped for sample collection. As soon as the
coupons came out of the HDP simulator they were soaked in isopropanol bath in order to
keep them from excessive oxidation.
The polished surfaces on the coupons were then identified, dried off and sheared down to
20 x 120 mm slices to further fit the GD-OES sample chamber. Next, the slices were taken
to GD-OES measurements, where a 4 mm-diameter rounded crater was carved on the
surface to sputter and collect material for quantitative chemistry surface analysis.
3.5.4. DUAL-STABILIZED IF STEEL GRADE
12 (twelve) dual-stabilized IF steel coupons were produced throughout the thermal profile
set up for dual-stabilized IF steels by interrupted cycles at the HDP simulator, one from
each spot where the thermal cycle was purposely stopped for sample collection. As soon as
the coupons came out of the HDP simulator they were soaked in isopropanol bath in order
to keep them from excessive oxidation.
70
The polished surfaces on the coupons were then identified, dried off and sheared down to
20 x 120 mm slices to further fit the GD-OES sample chamber. Next, the slices were taken
to GD-OES measurements, where a 4 mm-diameter rounded crater was carved on the
surface to sputter and collect material for quantitative chemistry surface analysis.
3.5.5. POST-PROCESSING AND STORAGE OF EXPERIMENTAL COUPONS
As soon as the coupons passed through the heat treatment cycles and came out of the last
cooling step of the thermal profile, they were soaked in an iso-propanol bath for the sake of
keeping them from excessive oxidation.
Afterwards, the samples remained soaked in iso-propanol up to the point when they were
taken out to be cut off down to 20 x 120 mm slices in order to fit the GD-OES sputtering
chamber.
3.6. SAMPLE ANALYSIS TECHNIQUES
IF steel bare samples were analyzed by using the following image and chemistry analysis
techniques: Scanning Electron Microscopy, Energy Dispersive Spectroscopy and Glow
Discharge Optical Emission Spectroscopy.
The samples with streaks on top of GA coatings were supplied by two different suppliers. In
order to establish comparisons between the areas featuring dark and light streaks, SEM
was used for image analysis and EDS for chemistry measurements, on a Philips SEM 515
Scanning Electron Microscope, attached to a Link Analytical Pentafet Energy Dispersive
Spectroscope micro analyzer.
SEM was also used as a tool for image analysis on samples with streaks on top of GA
coatings as well as its attached EDS microanalyzer for the chemistry measurements. A new
set of GD-OES profiles was run for the coated samples, this time aiming to identify the Fe-
Zn intermetallic phases, determine the thickness of the Γ layer and also obtain the Ti and
Nb concentration profiles through the GA coating, at the interfacial layer and into the steel
substrate.
71
GLOW DISCHARGE OPTICAL EMISSION SPECTROSCOPY FOR SAMPLE ANALYSIS
A Jobin-Yvon/Horiba Quantum IQ V2.22 Glow-Discharge Optical Emission Spectroscope
was used to obtain the surface chemistry profiles of the alloying elements for the as-
received samples.
X-RAY PHOTOELECTRON MICROSCOPY FOR SAMPLE ANALYSIS
The X-ray Photoelectron Spectroscopy PHI Quantera SXM was used to identify the
chemical state of Nb at the surface of IF steel substrates. The samples for the surface
characterization of Nb compounds at the surface of dual-stabilized IF steel grade were
taken from polished areas placed on the centre of the steel panels. The samples were then
sheared down to 15 x 15 mm and wrapped in aluminum foil in order to keep them from
external contamination. As the samples showed a slight C contamination when they were
placed in the XPS stage chamber due to the higher C diffusivity and segregation enthalpy
at lower temperatures (T < 400 0C) relative to Si and P for example, Ar ion bombardment
was used to sputter away the outer layers, therefore getting rid of the surface contaminants,
among those mainly C.
CRATER AND STREAK CHARACTERIZATION OF GA COATINGS
Industrial GA samples coating Ti-stabilized IF steel substrates featuring dark and light
streaked areas on top were received from steelmakers. Streak characterization was made
by both optical and electron microscopy.
Therefore, SEM was used as a tool to perform crater identification, characterization and
quantification. Furthermore, EDS analyzer was performed to assess the typical chemistries
found for craters on bottom of light and dark streaked regions.
CHARACTERIZATION OF STREAKS AND CRATERS ON TOP OF GA COATINGS FOR
INDUSTRIAL SAMPLES
Industrial GA samples with streaks on top were supplied by two different suppliers. The
samples were degreased with soap and ethanol was used to clean them for further drying
72
off. Streaked spots across the surface were tactfully chosen, which further lead to cut offs
for the Scanning Electron Microscope.
CHARACTERIZATION OF STREAKS AND CRATERS ON IF STEEL PANELS RUN
UNDER SPECIAL COOLING CONDITIONS AT THE HDP SIMULATOR
Ti- and dual-stabilized IF steel GA panels were selected for special tests at the HDP
simulator with the purpose of creating streaks on top of GA coatings. The thermal profiles
were set up taking into account the cooling conditions after the GA heat treatment. In order
to enable the operating conditions to create streaks, He gas was gushed in the fast cooling
chamber instead of 5%H2-N2 in order to speed up the cooling rate.
The thermal cycles featuring faster cooling rates for both IF steel grades can be seen on
Figure 3.
73
Figure 3. Thermal sequence for sampling with interrupted GA thermal cycles for both IF steel grades at the HDP simulator.
74
4. RESULTS AND DISCUSSION
4.1. MICROSTRUCTURE CHARACTERIZATION
The micrographs on Figures 4 and 5 show the typical microstructure found for the as-
received, i.e. as-cold-rolled samples, for both Ti- and dual-stabilized IF steel substrates.
The numbering of the Heyn Lineal Intercept Procedure for measuring the ferrite grain size is
also shown.
Figure 4. Micrograph featuring typical microstructures found for the as-received (as-cold-rolled) samples for Ti-stabilized IF steel substrate (optical microscope, magnification: 100X)
75
Figure 5. Micrograph featuring typical microstructure found for the as-received (as-cold-rolled) sample for dual-stabilized IF steel substrate (optical microscope, magnification: 100X).
The results carried out according to the Heyn Lineal Intercept Procedure for measuring the
ferrite grain size described on the ASTM E 112 - 96 standard are on table 7.
Table 7. Ferrite average grain sizes of the IF steels studied before and after galvannealing.
IF Steel Grade As-received Galvannealed
Ti-stabilized 21.74 µm 34.48 µm
Dual-stabilized 15.15 µm 21.74 µm
4.2. CHEMICAL PROFILES ON Ti- AND DUAL-STABILIZED IF STEEL
BARE AND COATED SAMPLES
The trial dipping samples ran at the HDP simulator in a 0.13 wt% effective Al in Zn bath and
same furnace controlling atmosphere, 95% N2 + 5% H2, showed that, for a Ti-only-
stabilized IF steel substrate, 8s at 510 0C were necessary to produce a fully-developed GA
coating featuring a 1 µm-thick Γ layer. On the other hand, 15s were needed for the dual-
76
stabilized IF steel substrate to produce the same results at the same temperature,
according to images shown on figure 6. Therefore, it can be concluded that a Ti-stabilized
IF steel substrate is more reactive than a dual-stabilized IF steel substrate.
Figure 6. SEM images at 2100x and 2020x featuring the GA coating cross section for the Ti-stabilized IF steel grade (left-hand side) and the dual-stabilized IF steel grade (right-hand side).
GD-OES was used as a tool to assess the surface elemental chemistry on bare steel
samples. As previously mentioned, 12 spots were chosen along the interrupted thermal
profile curve, where the samples were taken out of the HDG simulator and their surfaces
were assessed in order to understand the trend for both Ti and Nb surface concentrations
along the thermal profile.
Figures 7 and 8 feature the spots chosen along the thermal profiles for both IF steel grades
as well as the main areas of focus, i.e., the prior condition to sample dipping into the Zn
bath.
δ crystals
Γ layer
ζ crystals
77
Figure 7. Spots chosen along the Ti-stabilized IF steel thermal profile. Region within the solid red square feature the prior condition to sample dipping into the Zn bath.
Figure 8. Spots chosen along the dual-stabilized IF steel thermal profile. Region within the solid red square feature the prior condition to sample dipping into the Zn bath.
Therefore, as soon as the samples were taken out of the HDG simulator along the
interrupted thermal profiles mentioned above, their surface chemistries were assessed
through GD-OES analysis.
78
Figures 9 to 12 show the Ti and Nb outermost surface concentration profiles along the
thermal curves for both Ti- and dual-stabilized IF steel substrates found out for bare
samples.
Figure 9. Ti outermost surface concentration along the spots where bare steel samples were taken out of Ti-stabilized IF steel thermal profiles. (S1 to 12 => temperature vs. time condition)
79
Figure 10. Nb outermost surface concentration along the spots where bare steel samples were taken out of Ti-stabilized IF steel thermal profiles. (S1 to 12 => temperature vs. time condition)
80
Figure 11. Ti outermost surface concentration along the spots where bare steel samples were taken out of dual-stabilized IF steel thermal profiles. (S1 to 12 => temperature vs. time condition)
81
Figure 12. Nb outermost surface concentration along the spots where bare steel samples were taken out of dual-stabilized IF steel thermal profiles. (S1 to 12 => temperature vs. time condition)
Therefore, GD-OES results for bare samples on both IF steel grades showed that there is a
Ti depletion at the outermost surface of both Ti- and dual-stabilized IF steel substrates. The
influence of annealing and galvannealing heat treatments upon Ti surface segregation is
82
minimal and not remarkable. Up to a sputtering depth of 0.02 µm into the steel substrate, a
Ti depletion was found on most samples along the thermal profile for both IF steel grades.
Moving farther into the steel substrate - 0.1 µm – a slight Ti enrichment was found for most
part of samples on both IF steel substrates. This enrichment is likely to be associated with a
small amount of Ti that has not been previously tied up, therefore available in solid solution.
According to excess Ti calculations based on the formula excess Ti = total Ti – 3.99C –
1.49S – 3.42N – 1.55P given by Marder et al., there are 90 ppm Ti left in solid solution and
available for diffusion prior to the annealing heat treatment on the Ti-only-stabilized IF steel
grade. GD-OES chemistry results for Ti can be seen on figure 13.
Figure 13. Ti surface concentration results at 0.02 µm into the Ti- and dual-stabilized IF steel substrates along the thermal cycles set up for both IF steel grades (Ti bulk concentrations are 0.062 at% and 0.024 at% respectively).
Ti Content at 0.02 um into the Dual-stabilized IF Steel Substrate
0
0,005
0,01
0,015
0,02
0,025
0 50 100 150 200 250 300 350Heat Treatment Time (s)
Co
nce
ntra
tion
(at%
)
83
On the other hand, a strong Nb enrichment was found at the nearest surface on both Ti-
and dual-stabilized IF steel substrates. As soon as the samples started to be heated
upwards the annealing peak, part of Nb in solid solution starts to diffuse towards the
outermost surface of the steel substrate, driven by the tendency of chemical potential
uniformity for this element between the bulk of the steel substrate and the surface. In the
wake of the annealing heat treatment, those high Nb concentrations remain at the nearest
surface all the way through the end of the GA heat treatment cycle. Besides, during
galvannealing heat treatment no remarkable effect has been observed upon the Nb surface
segregation. Nb surface concentration profiles are featured on figure 14.
Figure 14. Nb surface concentration results at 0.02 µm into the Ti- and dual-stabilized IF steel substrates along the thermal cycles set up for both IF steel grades (Nb bulk concentrations are 0.005 at% and 0.029 at%, respectively).
Nb Content at 0.02 um into the Ti-stabilized IF Steel Substrate
0
0,005
0,01
0,015
0,02
0,025
0,03
0 50 100 150 200 250 300 350 400 450Heat Treatment Time (s)
Co
nce
ntr
atio
n (
at%
)
84
XPS analyses were carried out aiming to assess the Nb chemical state for the as-
galvannealed condition at the surface of IF steels. For this purpose, the samples chosen for
such analysis were those taken at the threshold of the GA heat treatment, which represent
the setting condition for the subsequent galvannealing heat treatment. The XPS results
showed a characteristic Nb 3d3/2 peak, which corresponds to Nb oxides in the form of
Nb2O5, as can be seen on figure 15.
Figure 15. XPS surface analysis results for the dual-stabilized IF steel grade, featuring a Nb 3d3/2 peak whose binding energy is 207.4 eV), which corresponds to Nb2O5.
Therefore, it is possible that one of the reasons why the Ti-stabilized IF steel substrate is
more reactive than the dual-stabilized IF steel substrate is that there are Nb oxides at the
Zn coating-steel substrate interface blocking the Fe-Zn interdiffusion along the ferrite grain
boundaries on the dual-stabilized IF steel substrate, whose Nb content is much higher than
that for the Ti-only-stabilized IF steel substrate.
Regarding the setting up for GA heat treatment profiles for both IF steel grades, the GA
heat treatment temperature chosen for the HDG simulator experiments was 510 0C.
However, as the IF steel substrate reactivity depends on the Ti and Nb bulk concentrations,
experiments were carried out in order to attain the GA heat treatment time necessary to
produce a 1 µm-thick Γ layer on top of both IF steel substrates.
2002052102150
100
200
300
400
500
Binding Energy (eV)
c/s
Nb3d
2002052102150
100
200
300
400
500
Binding Energy (eV)
c/s
Nb3d
3d3/2
3d1/2
85
As indicated by the mainstream literature on IF steels, Ti-stabilized IF steel substrates are
more reactive than dual-stabilized IF steel substrates. Therefore, the time necessary to
produce a 1 µm-thick G layer on top of a Ti-stabilized IF steel substrate (8 s) is shorter than
for dual-stabilized IF steel substrates (15 s).
Figures below feature the GD-OES Fe, Zn and Al elemental concentrations across the GA
coating for both IF steel substrates after GA heat treatment at 510 0C and GA heat
treatment times mentioned above, according to the IF steel substrate.
Figure 16. GD-OES Zn, Fe and Al atomic concentration across GA coating for a Ti-stabilized IF steel substrate.
86
Figure 17. GD-OES Zn, Fe and Al atomic concentration across GA coating for a dual-stabilized IF steel substrate.
Even though the GA coating thicknesses observed for Ti- and dual-stabilized IF steels had
not been uniform throughout the HDG simulation experiments, it is worth pointing out that
the purpose of this project does not include setting up a uniform GA coating thickness,
especially as it has neither straight correlation with IF steel substrate reactivities nor
cratering formation mechanisms.
Regarding the surface concentration for alloying elements Ti and Nb across GA coatings for
both IF steels, GD-OES experiments carried out on top of GA coated samples pointed out
that both Ti outermost concentrations as well as Ti concentrations at the GA coating-steel
substrate interface are significantly lower than those found at the bulk steel substrate,
regardless of the IF steel substrate.
On the other hand, GD-OES elemental analysis for Nb have shown that both Nb outermost
concentration as well as Nb concentrations at the GA coating-steel substrate interface are
much higher than those found at the respective bulk steel substrate, regardless of the IF
steel substrate.
Figures 18 and 19 feature the alloying element concentrations along the GA coating for
both IF steel substrates.
87
Figure 18. Ti and Nb concentrations along a 30 µm-thick GA coating on top of a Ti-stabilized IF steel substrate.
88
Figure 19. Ti and Nb concentrations along a 12 µm-thick GA coating on top of a dual-stabilized IF steel substrate.
4.3. CRATER ANALYSIS ON TOP OF GA COATINGS
4.3.1. CRATER ANALYSIS ON LIGHT AND DARK STREAKED AREAS FOR AS-
RECEIVED GA COATINGS
Both light stereoscope and optical microscope micrographs were taken from the samples
with streaks on top of GA coating. Two distinct types of streaks were found, a dark and a
light one, which owe their distinctive features to differences in reflectivity and chemistry
non-uniformity on top of the GA coating.
Figure 20 shows light stereoscope pictures featuring macro overviews of streaked regions
on top of GA coatings.
89
Figure 20. Stereoscope micrographs featuring light and dark streaked areas on top of GA coatings for Ti-stabilized IF steels. Magnification: 6.5 x.
After that, SEM image analysis was used to magnify the region on dark and light streaked
areas, as can be seen on figure 21.
Figure 21. SEM micrographs featuring light (right-hand side) and dark (left-hand side) streaked areas on top of GA coatings for Ti-stabilized IF steels. Magnification: 573 x.
Furthermore, zooming in the SEM magnification within the light and dark streak areas,
craters were found on the bottom of these regions. The areas featuring dark streaked areas
showed a higher crater coverage when compared with the light streaked regions. EDS
chemistry measurements were then carried out and showed that δ crystals were found on
the bottom of craters on light streaked areas, whereas there is Γ layer on the bottom of
craters on dark streaked areas. Therefore, the craters on dark streaked regions are deeper
90
than those on light streaked areas. Micrographs and typical chemistries found for both kinds
of craters are shown on figure 22.
Element wt% at%Al 1.91 4.35Fe 21.39 23.54Zn 76.70 72.11
Figure 22. SEM images at 2300x featuring typical microstructures for a dark streaked area (left-hand side), whose EDS chemistry taken from the bottom matches a Γ phase, and light streaked area (right-hand side), whose chemistry is typical of a δ phase.
4.3.2. SAMPLES RUN UNDER SPECIAL COOLING CONDITIONS AT THE HOT-DIP
PROCESS SIMULATOR
Regarding the Ti- and dual-stabilized IF steel panels that run interrupted thermal cycles
under special cooling conditions at the HDP simulator in order to reproduce the operating
conditions for crater formation in a CGL, further SEM analysis and EDS chemistry results
showed that craters were only found on top of GA coatings for Ti-stabilized IF steel
substrates, no matter if the panels were heat treated for 0s, 2s or 4s. On the other hand, no
craters were found for dual-stabilized IF steel samples, regardless of the GA heat treatment
time. Therefore, contending the previous work by most part of authors, Ti-stabilized IF steel
substrates are more prone to crater formation than dual-stabilized IF steel substrates.
Figure 23 shows the overall GA coating surface for the panels run for both IF steel grades.
Element wt% at%Al 0.60 1.41Fe 11.39 12.97Zn 88.02 85.62
91
Figure 23. Pictures feature the overall look of Ti- (left-hand side) and dual-stabilized (right-hand side) IF steel panels that run interrupted GA thermal cycles under faster cooling rates.
Moreover, Ti-stabilized IF steel panels featuring light and dark streaked areas on top of
their GA coating were taken to the Scanning Electron Microscope for the purpose of streak
and crater characterization through SEM micrographs and EDS chemistry, whose results
are shown on figures 24 and 25.
Figure 24. SEM images and EDS chemistries for craters on bottom of light streaked areas at GA coatings on top of Ti-stabilized IF steel panels after running interrupted GA thermal cycles under faster cooling rates.
Panel Sampled after 2s GA HeatTreatment Time
SEM Image – 2020x
Element Concentration(wt%)
Concentration(at%)
Al 2.11 4.89Fe 9.28 10.39Zn 88.61 84.72
Panel Sampled after 4s GA HeatTreatment Time
SEM Image – 2020x
Element Concentration(wt%)
Concentration(at%)
Al 0.25 0.60Fe 10.89 12.47Zn 88.85 86.92
92
Figure 25. SEM images and EDS chemistries for craters on bottom of dark streaked areas at GA coatings on top of Ti-stabilized IF steel panels after running interrupted GA thermal cycles under faster cooling rates.
The EDS chemistries found at the bottom of the craters proved that d crystals are
connected with light streaked areas whereas G phase is linked with dark streaked regions.
4.4. STEEL SUBSTRATE REACTIVITY ALONG FERRITE GRAIN
BOUNDARIES ON Ti- AND DUAL-STABILIZED IF STEEL SUBSTRATES:
According to the literature, Ti-bearing precipitates, mainly TiN, segregate to regions far from
the ferrite grain boundaries, for the reasons explained on previous sessions. Therefore,
those compounds are responsible for “cleaning” the ferrite grain boundaries for speeding up
further Fe and Zn interdiffusion along the ferrite grain boundaries by the time the steel sheet
is immersed into the Zn galvanizing bath.
On the other hand, Nb-bearing precipitates, mainly NbC compounds, segregate to ferrite
grain boundaries. Thus the Nb compounds get in the way of Fe and Zn interdiffusion along
the grain boundaries as the steel sheet is immersed into the Zn galvanizing bath.
The results obtained by this research agree with the statements above. The galvanneal
heat treatment time at 510°C necessary to end up with a 1 µm-thick G layer on top of the
Panel Sampled after 2s GA HeatTreatment Time
SEM Image – 2020x
Element Concentration(wt%)
Concentration(at%)
Al 0.68 1.61Fe 12.05 13.69Zn 87.27 84.70
Panel Sampled after 4s GA HeatTreatment Time
SEM Image – 2020x
Element Concentration(wt%)
Concentration(at%)
Al 0.29 0.67Fe 19.14 21.61Zn 80.57 77.72
93
steel substrate was longer for the dual-stabilized IF steel substrate when compared with
shorter heat treatment times needed for Ti-stabilized IF steel substrate.
4.4.1. REGIONS OF HIGH REACTIVITY AND FORMATION OF OUTBURSTS
The absence of Ti precipitates, mainly TiN compounds, at the ferrite grain boundaries mean
that the Ti-stabilized IF steel substrates are more reactive than the dual-stabilized IF steel
substrates.
Therefore, Ti-stabilized IF steel substrates are more prone to outburst formation than dual-
stabilized IF steel grades due to the faster Fe and Zn interdiffusion right after the steel
sheet enters the Zn bath.
Furthermore, literature reports that craters are clusters of outbursts on top of the GA
coatings, which blocks the nucleation and growth of Fe-Zn intermetallic phases on top of
the steel substrate. Indeed, no craters were found amid the GA coatings on top of dual-
stabilized IF steel panels, regardless of the GA heat treatment time.
4.4.2. REGIONS OF HIGH REACTIVITY AND FORMATION OF CRATERS ON LIGHT
AND DARK STREAKED AREAS
Literature reports that the brightness differences on cratered regions on top of GA coatings
are due to differences in reflectivity.
Indeed, the SEM as well as EDS results showed that there are two different kinds of craters
on bottom of GA coatings, as follows: dark craters due to the presence of G layer on their
bottom; light craters due to the presence of d crystals on their bottom.
Therefore, the research results agree with the literature.
94
4.5. THE EFFECT OF THE DIFFUSION OF IF STEEL ALLOYING
ELEMENTS TOWARDS THE INTERFACE BETWEEN IF STEEL
SUBSTRATES AND GA COATINGS ON CRATER FORMATION
The experimental results presented above have shown that IF steel alloying elements,
namely Ti and Nb, feature distinctive characteristics at the GA coating-steel substrate
interface.
The Ti and Nb free atoms in solid solution scattered across the typical ferrite microstructure
of IF steels are bound to segregate as soon as the temperature is risen, therefore
thermodynamics as well as kinetics play an important role when it comes to GA coating-
steel substrate alloying element segregation.
Finally, based on the bulk chemistry given for each IF steel grade, useful formulas are
frequently used in order to estimate the final free amount of Ti and Nb left in solid solution,
taking into account the C, P, N and S contents at the bulk of the respective steel substrate.
4.5.1. Ti DEPLETION AT THE STEEL SUBSTRATE-GA COATING INTERFACE;
The experimental procedures carried out by this research have shown that the outermost
surface of the steel substrates are free from Ti-bearing compounds shortly before the strip
enters the Zn pot.
The statement above has been proved by suddenly stopping the thermal cycle at the HDG
simulator and taking the steel coupon at the spot whose temperature and time set the
conditions for Zn bath entrance for both IF steel substrates i.e. the surface condition upon
which the steel substrate is submitted to Zn wetting.
As mentioned before, GD-OES elemental analysis were carried out for bare IF steel
samples, and those representing the setting condition for the Zn bath immersion featured
no Ti contents whatsoever above the average Ti content found in the bulk of the steel
substrate.
95
Therefore, the Ti depletion at the surface of both IF steel substrate means that Ti-bearing
compounds have no influence whatsoever on the Fe-Zn interdiffusion rates along the
interface between the steel substrate and the GA coating, even for short circuit paths for
diffusion i.e. ferrite grain boundaries.
4.5.2. Nb ENRICHMENT AT THE STEEL SUBSTRATE-GA COATING INTERFACE
The experimental procedures carried out by this research have shown that the outermost
surface for both IF steel substrates at the time of strip entrance into the Zn bath feature a
Nb concentration significantly higher than the Nb contents found in the bulk of both IF steel
substrates.
The statement above has been proved by suddenly stopping the thermal cycle at the HDG
simulator and taking the steel coupon at the spot whose temperature and time set the
conditions for Zn bath entrance for both IF steel substrates i.e. the surface condition upon
which the steel substrate is submitted to Zn wetting.
As mentioned before, GD-OES elemental analyses were carried out for bare IF steel
samples, and those representing the setting condition for the Zn bath immersion featured
higher Nb concentrations than those found at the bulk of both IF steel substrates.
Furthermore, XPS analysis have shown that the Nb-bearing compounds at the surface of
steel coupons taken from the HDG simulator correspond to Nb oxides in the form of Nb2O5.
Therefore, the Nb enrichment at the surface of both IF steel substrate means that it affects
the steel substrate reactivity, thus Nb oxides get in the way of Fe-Zn interdiffusion rates
along the interface between the steel substrate and the GA coating by slowing down the Fe
and Zn diffusivities across the interface.
Moreover, as typical bulk Nb contents are much higher for dual-stabilized IF steel grades
when compared with those for Ti-stabilized IF steel grades, the effect of Nb surface
96
enrichments upon the reactivity of dual-stabilized IF steel grades is much improved when
compared with Ti-stabilized IF steel substrates.
Indeed, the longer GA heat treatment times necessary to produce a 1 µm-thick G layer on
top of dual-stabilized IF steel substrates when compared with the shorter ones necessary
for Ti-stabilized IF steel substrates are enough to prove the statement above.
4.5.2.1. MECHANISMS FOR DRIVING Nb OXIDATION AT THE STEEL SUBSTRATE-GA
COATING INTERFACE.
After cold-rolling to final thickness, steel sheets need to be recrystallized in order to recover
intrinsic ductility for cold forming and drawing applications. This annealing treatment is
performed either by batch-annealing of tight coils for long soaking times at intermediate
temperatures or by continuous annealing of steel strip at higher temperature for shorter
times, the later being the case carried out by this Master’s thesis.
Therefore, a protective atmosphere based on either N2-H2 mixed gas or pure hydrogen is
present in the furnace. This atmosphere is surely reducing for Fe but very often oxidizing for
more reactive alloying or residual elements present in the steel, depending on the residual
oxygen content and dew point of the concerned protective atmosphere.
In such conditions, it is anticipated that selective oxidation of alloying elements will take
place at the steel surface producing extremely different surface chemistries for steel sheets,
what results in different chemical reactivities in subsequent finishing treatments, for
instance Zn coating by hot-dipping.
Wagner et al.68 have put forward the theoretical basic principles of a mathematical model
describing selective oxidation in ideal single crystal binary alloys where it is assumed that
selective oxidation is governed by the interdiffusion of reactive species.
Moreover, both the localization of the reaction and precipitation front relative to the free
surface depends on the balance between the inward oxygen and outward oxidizable solute
97
element flows. For well balanced flows, reaction as well as oxide precipitation are observed
beneath the free surface and selective oxidation is called internal.
On the other hand, external oxidation can be treated like a particular case of the internal
mode, when selective surface oxide precipitation, resulting from inward and outward flow
balance is predominant enough as to obstruct all oxygen diffusion paths.
However, the complexity of the steel chemistries and the possibility to form complex or
mixed oxides made the use of the Wagner’s model quite difficult. For instance, the Wagner
criterion cannot account for the grain boundary oxidation process observed in specific
cases. Indeed, it is proposed to take account of selective oxidation at grain boundary by
rewriting the Wagner criterion in such a form that the accelerated diffusion flows of oxygen
and alloying element along these grain boundaries are considered.
However, due to the complex chemistry of steels, predictions based on relations derived
from Wagner’s criterion applied to bulk or grain boundary diffusion are generally far from
accurate.
Regarding the Nb selective oxidation, it is considered by Birks and Meier69 as a system with
significant scale cracking. The high temperature oxidation of Nb is characterized by inward
diffusion of oxygen through the scale. Initially, a protective layer is formed but, as the scale
grows, the formation of oxide at the scale-metal interface stresses the oxide resulting in
scale cracking and a breakaway linear oxidation.
Indeed, finding out the driving mechanisms for surface Nb oxide formation has not been the
concern of this Master’s thesis. As mentioned earlier XPS analysis on the bare surface for
dual-stabilized IF steel substrates whose samples were taken at setting conditions for hot-
dipping into the Zn bath has found Nb-bearing compounds namely Nb2O5.
Further recommendations for the follow up of this work include the following issues to be
addressed focusing on further assessing the Nb oxide formation mechanisms:
98
- Transmission Electron Microscopy as a tool to spot the right location of Nb oxides on top
of the dual-stabilized IF steel samples i.e. whether they are placed at the ferrite grain
boundary intersections or along the bulk of the ferrite grains;
- Attaining a mathematical model in order to assess the mechanisms upon which free Nb
has been driven outwards and tying up with oxygen as well as the thermodynamics
(temperature and pressure conditions) behind the driving force for such mechanism.
4.5.3. EFFECT OF Ti AND Nb SURFACE SEGREGATION ON CRATER FORMATION
Literature in general has stated that craters on top of galvanneal coatings are formed due to
the presence of liquid Zn close to ? crystals right after the strip entrance to the Zn bath.
Provided the reactivity of the steel substrate is quite high, i.e. high Fe-Zn interdiffusivity
along the ferrite grain boundaries, the liquid Zn is drained towards the outbursts formed at
the GA coating-steel substrate, thus creating the streaks, which are clusters of craters on
top of GA coatings.
Furthermore, some authors have found that high Ti contents at the IF steel substrates are
prone to streak formation, thus proving the statement above.
The first task of this research toward craters was to provide its thorough characterization
through optical microscope analysis, SEM microanalysis as well as EDS chemistry analysis
on industrial samples taken from Ti-stabilized IF steel substrates.
Sample microanalysis carried out through optical microscope found that there are two basic
types of streaks, i.e. cluster of craters on top of GA coatings, light and dark streaked areas.
Moreover, SEM analysis proved that the craters found on the bottom of light streaked areas
are different than those found on the bottom of dark streaked areas when it comes to crater
density and chemistry.
99
Therefore, the average crater density for dark streaked areas is significantly higher than for
light streaked areas. Additionally, the EDS analysis carried out for both types of craters
concluded that the chemistry for the bottom of craters on light streaked areas is similar to d
Fe-Zn intermetallic phase, whereas the chemistry for the bottom of dark streaked areas is
similar to G Fe-Zn intermetallic phase.
Having said that, we can now point out that craters on dark streaked areas are deeper than
those found on light streaked areas.
In order to replicate the industrial condition at the HDG simulator, bare samples were
spared for both IF steel substrates, taking into account the thermal profiles set up thus far.
However, the GA heat treatment was suddenly interrupted at three different spots and then
cooling rates right after GA heat treatment were changed in order to enhance the kinetics
towards crater formation by using He gas in a special cooling chamber.
Results from this experimental procedure pointed out that craters were only found at GA
coatings on top of Ti-stabilized IF steel substrates regardless of the GA heat treatment time.
On the other hand, no craters were found at GA coatings on top of dual-stabilized IF steel
substrates regardless of the GA heat treatment time.
Therefore, as agreed with the literature in general, Ti content indeed plays an important role
on streak and crater formation on top of GA coatings.
100
5. CONCLUSIONS
Ti depletion was found at the outermost surface on both IF steel grades, whereas this
region features a strong Nb enrichment.
According to XPS analysis, this Nb surface enrichment corresponds to Nb oxides (Nb2O5).
The presence of Nb oxides at the surface of dual-stabilized IF steel substrates may block
the Fe-Zn interdiffusion along the ferrite grain boundaries.
Simulations under special cooling conditions at the Hot-Dip Process Simulator found
streaks only on Ti-stabilized IF steel substrate.
There is higher crater density on dark streaked areas when compared with light streaked
regions of IF galvannealed steels.
δ crystals were found on the bottom of craters on light streaked areas, whereas there is Γ
phase on the bottom of craters on dark streaked areas. Also, the craters on light streaked
areas are shallower than those on dark streaked areas.
Due to the presence of Nb oxides at the interface between the Zn coating and the dual-
stabilized IF steel substrate, the Ti-only-stabilized IF steel substrate is more reactive and
consequently more prone than the dual-stabilized IF steel substrate to the formation of
outbursts, and consequently craters.
101
6. ACKNOWLEDGEMENTS
We would just like to say many thanks to each and every person who have been in
some way connected with the experimental procedures as well as the results presented on
this research project. Among them it is noteworthy to include the sponsors for the Master's
project whose framework this work comes from i.e. CBMM as well as ArcelorMittal
Tubarão’s Customer Service and HR personnel, especially for the necessary support
throughout my two-year stint as a Master’s student at McMaster University in Canada.
Besides them, special acknowledgments must be made to the supervisors whose valuable
coaching has been paramount towards the soundness of the results attained, as such Dr
Fernando Gabriel Araújo at Redemat/UFOP as well as Dr Joseph McDermid at McMaster
University, also including MSE faculty, staff and alumni.
102
7. BIBLIOGRAPHY
1. Bhattacharya, D. and Cheng, C. Mechanism of the Effect of Nb on the Galvannealing
Behavior of IF Steels. Proceedings of the 6th International Conference on Zinc and Zinc
Alloy Coated Steel Sheet (GALVATECH’04), pp. 509-516, Chicago, 2004
2. Jordan, C. E. and Marder, A. R. Effect of Substrate Grain Size on Iron-Zinc Reaction
Kinetics during Hot-Dip Galvanizing. Metallurgical and Materials Transactions A, volume
28A, December 1997, pp. 2683-2694, 1997
3. Osman, T. M. and Garcia, I. Niobium-Bearing Interstitial-Free Steels: Processing,
Structure and Properties. Pittsburgh, USA: Basic Metals Processing, Research Institute
4. Feliu, S. and Perez-Revenga, M. L. Effect of the Presence of Alloying Elements in
Interstitial-Free and Low-Carbon Steels on their Surface Composition after Annealing in
Reducing Atmospheres (Dew Point = -300C). Madrid, Spain: CENIM (CSIC), 2003
5. Edwards, C.A.; Phillips, D. L.; Jones, H. N.; The Influence of Some Special Elements
Upon the Strain-Ageing and Yield-Point Characteristics of Low-Carbon Steels. J. Iron and
Steel Institute, 142, pp. 199-236. 1940.
6. Codd, I.; Petch, N.J. Dislocation-Locking by Carbon, Nitrogen and Boron in a-Iron. Phil.
Mag. 5, pp. 30-42. 1960
7. Hook, R.E. The Effect of Quench-Ageing on Inhomogeneous Yielding of Steels with Very
Low Interstitial Solute Content. Metallurgical Transactions, pp. 85-92. 1970.
8. Marder, A. R. The Metallurgy of Zinc-Coated Steel. Progress in Materials Science 45
(2000) pp. 191-271, 2000
103
9. Lin, C. S. and Meshii, M. The Effect of Steel Chemistry on the Formation of Fe-Zn
Intermetallic Compounds of Galvanneal-Coated Steel Sheets. Metallurgical and Materials
Transactions B, volume 25B, October 1994, pp. 721-729, 1994
10. Baril, E. and L’Esperance, G. Studies of the Morphology of the Al-Rich Interfacial Layer
Formed During the Hot Dip Galvanizing of Steel Sheet. Metallurgical and Materials
Transactions A, volume 30A, March 1999, pp. 681-695, 1999
11. Parks, W. H.; Haggerty, C. S. and Rock, T. R. Ferritic Rolling of Interstitial-Free Steel.
Iron and Steel Engineer, October 1997, pp. 35-36, 1997
12. van der Heiden, A.; Burghardt, A. J. C.; van Koesveld, W.; van Perlstein, E. B. and
Spanjers, M. G. J. Galvanneal Microstructure and Anti-Powdering Process Windows. The
Physical Metallurgy of Zinc Coated Steel, pp. 251-263, 1993
13. McDevitt, E. T. and Meshii, M. Microstructural Evolution During Galvannealing. Part II:
Formation and Growth of the Fe-Zn Intermetallic Compounds. Proceedings of the 4th
International Conference on Zinc and Zinc Alloy Coated Steel Sheet (GALVATECH’98),
Chiba, Japan, The Iron and Steel Institute of Japan, 1998
14. McDevitt, E. T.; Morimoto, Y. and Meshii, M. Microstructural Evolution During
Galvannealing. Part 1: Formation and Breakdown of the Fe-Al Inhibition Layer. Chiba,
Japan: The 4th International Conference on Zinc and Zinc Alloy Coated Steel Sheet
(GALVATECH’98). The Iron and Steel Institute of Japan, 1998
15. McDevitt, E. T.; Pelayo, L; Morimoto, Y. and Meshii, M. Investigation of the Fe-Al
Inhibition Layer Formed During Hot-Dip Galvannealing. 39th MWSP Conference
Proceedings, ISS, vol. XXXV, pp. 127-135, 1998
16. Dionne, S.; Charest, M.; Botton, G. and Goodwin, F. Influence of Substrate
Characteristics on the Inhibition Layer Breakdown and Galvannealing Kinetics. 44th MWSP
Conference Proceedings, vol. XV, pp. 405-415, 2002
104
17. Jordan, C. E. and Marder, A. R. Alloy Layer Growth During Hot-Dip Galvanizing at
4500C. Proceedings of The 5th International Conference on Zinc and Zinc Alloy Coated
Steel Sheet (GALVATECH’95), pp. 319-325, Chicago, USA, 1995.
18. McDevitt, E. T.; Morimoto, Y. and Meshii, M. Microstructural Characterization of the Fe-
Al Inhibition Layer on Commercial Hot-Dip Galvanized IF Steel. 39th MWSP Conference
Proceedings, ISS, vol. XXXV, pp. 215-223, 1998
19. O’Dell, S. P.; Charles, J. and Randle, V. The Formation of Craters Within the
Galvanneal Coating and their Effects on Processing Properties. 44th MWSP Conference
Proceedings, vol. XL, pp. 1299-1307, 2002
20. Lee, H. J. and Kim, J. S. Effect of Ni Addition in Zn Bath on Formation of Inhibition
Layer During Galvannealing of Hot-Dip Galvanized Sheet Steels. Journal of Materials
Science Letters 20, pp. 955-957, 2001
21. McDermid. J. R.; Baril, E. and Thompson, W. T. Fe Solubility in the Zn-Al-Fe System for
Use in Continuous Galvanizing and Galvannealing. Hamilton, ON, Canada: Department of
Mechanical Engineering, McMaster University
22. Miyasaka, A. and Matsumura, K. Influence of Alloying Elements in Steel on
Galvannealing Behavior of Interstitial-Free (IF) Steels. Proceedings of The 4th International
Conference on Zinc and Zinc Alloy Coated Steel Sheet (GALVATECH’98), pp. 186-189,
Chiba, Japan, The Iron and Steel Institute of Japan, 1998.
23. Jordan, C. E.; Zuhr, R. and Marder, A. R. Effect of Phosphorus Surface Segregation on
Iron-Zinc Reaction Kinetics during Hot-Dip Galvanizing. Metallurgical and Materials
Transactions A, volume 28A, December 1997, pp. 2695-2703, 1997
24. Perrot, P; Tissier, J. C. and Dauphin, J. Y. Stable and Metastable Equilibria in the Fe-
Zn-Al System at 4500C. Villeneuve d’Ascq, Cedex, France: Laboratoire de Metallurgie
Physique, Universite de Lille, 1992
105
25. Shindo, Y. Effect of Different Alloying Elements in Steel on Galvanneal Coating
Formation. Ispat Inland Research and Development
26. Maschek, W.; Hayes, S. P. and Marder, A. R. Cross Sectional Studies of Zinc Iron
Phase Growth in an Environmental Scanning Electron Microscope. Proceedings of The 5th
International Conference on Zinc and Zinc Alloy Coated Steel Sheet (GALVATECH’95),
Chicago, USA, 1995.
27. Zhong, W.; Ng, H. F. and James, J. M. Correlation Between Adhesion Properties and
the Interfacial Bonding Strength of Galvanneal Coatings. Proceedings of the 4th
International Conference on Zinc and Zinc Alloy Coated Steel Sheet (GALVATECH’98),
Chiba, Japan, The Iron and Steel Institute of Japan, 1998
28. Angeli, G.; Deinhammer, H.; Faderl, J.; Angeli, J.; Papst, I. and Wolpers, M. Distribution
and Composition of Surface Oxides on High Strength IF-Steels after Recrystallization
Annealing. 44th MWSP Conference Proceedings, vol. XL, pp. 825-832, 2002
29. Feliu, S.; Perez-Revenga, M. L. and Barranco, V. Persistence of Superficial
Contamination of Rolled Steel during Successive Treatments up to the Formation of a
Galvanneal Coating. Metallurgical and Materials Transactions A, volume 35A, July 2004,
pp. 2185-2187, 2004
30. Faderl, J.; Strutzenberger, J. and Fischer, W. Galvannealed Steel Sheet: 10 Years of
Experience on Product and Process Improvement. Linz, Austria: Voest Alpine Stahl
31. van Koesveld, W.; Lamberigts, A.; van der Heiden, A. and Bordignon, L. Coating
Microstructure Assessment and Control for Advanced Product Properties of Galvannealed
IF Steels. Proceedings of the 3rd International Conference on Zinc and Zinc Alloy Coated
Steel Sheet (GALVATECH’95), pp. 343-354, 1995
32. Hashiguchi, K.; Yasuda, A.; Hanazawa, T; Ohori, M. and Ichida, T. Galvannealed Sheet
Steel with Excellent Press Formability. La Revue de Métallurgie – CIT, Mars 1990, pp. 277-
283, 1990
106
33. Miner, R. E. and Butler, J. F. Comparison of Product Characteristics of
Electrogalvanized and Galvannealed for Exposed Automotive Applications. Proceedings of
The 5th International Conference on Zinc and Zinc Alloy Coated Steel Sheet
(GALVATECH’95), pp. 407-412, Chicago, USA, 1995.
34. Elias, C. N.; Alcala-Vela, J. and Maestri, C. Microstructural Analysis of Galvannealed
Coating by Grazing-Incidence X-Ray Diffraction. Volta Redonda, RJ, Brazil: Escola de
Engenharia Metalúrgica de Volta Redonda
35. Ebrill, N.; Durandet, Y. and Strezov, L. Dynamic Wetting and its Influence on Interfacial
Resistance during Hot Dip Galvanizing. Shortland, Australia: BHP Minerals Technology
36. Zhong, W.; Ng, H. F. and Ichikawa, M. Effect of Coating Characteristics on Powdering
Performance of Galvanneal Coatings. Proceedings of the International Symposium held at
the TMS Annual Meeting, San Antonio, Texas, 1998
37. Wilmotte, S.; Fischbach, J. P. and Hardy, Y. Control of the Galvannealing Process.
CRM-Belgium, SEGAL-Belgium
38. Lamberigts, M.; Beguin, M.; Dubois, M.; van der Heiden, A. and Verhoeven, G. Kinetics
of Fe-Zn Reactions During Galvannealing. CRM, Segal, Cockerill Sambre, Hoogovens
Groep B.V, N. V. Sidmar. Belgium. The Netherlands
39. Divinski, S. V.; Geise, J.; Rabkin, E. and Herzig, Chr. Grain Boundary Self-Diffusion in
α-Iron of Different Purity: Effect of Dislocation Enhanced Diffusion. Z. Metallkd. 95 (2004)
10, pp. 945-952, 2004
40. Faderl, J.; Pimminger, M. and Schönberger, L. Influence of Steel Grade and Surface
Topography on the Galvanneal Reaction. Voest-Alpine Stahl Linz GmbH
107
41. Drillet, P.; Dulcy, C.; Pazsko, F.; Mangelinck, D.; Gas, P.; Clugnet, G.; Bergman, C. and
Vaughan, G. Real Time Synchrotron Analysis of the Initial Stages of the Galvanization
Process in Al Containing Zn Baths. La Revue de Métallurgie-CIT, Janvier 2005, pp. 75-78
42. Rege, J. S.; Hua, M.; Garcia, C. I. and DeArdo, A. J. The Segregation Behavior of
Phosphorus in Ti and Ti + Nb Stabilized Interstitial-Free Steels. ISIJ International, vol. 40
(2000), no. 2, pp. 191-199, 2000
43. Nassar, A.; Vipond, R. and Mintz, B. Influence of Intercritical Annealing on Strength and
Impact Behavior of Niobium Containing Steels. Materials Science and Technology, August
1991, vol. 7, pp. 699-706, 1991
44. Alaoua, D.; Larere, A.; Lartigue, S. and Priester, L. Interfacial Segregation in Ultra-Low
Carbon Steels. Materials Science Forum vol. 126-128 (1993), pp. 185-188, 1993
45. Olefjord, I.; Leijon, W. and Jelvestam, U. Selective Surface Oxidation During Annealing
of Steels Sheets in H2/N2. Applications of Surface Science 6 (1980) pp. 241-255, 1980
46. Alaoua, D.; Lartigue, S.; Larere, A. and Priester, L. Precipitation and Surface
Segregation in Low Carbon Steels. Materials Science and Engineering, A189 (1994), pp.
155- 163, 1994
47. Hua, M.; Garcia, C. I. and DeArdo, A. J. Precipitation Behavior in Ultra-Low-Carbon
Steels Containing Titanium and Niobium. Metallurgical and Materials Transactions A,
volume 28A, September 1997, pp. 1769-1780, 1997
48. Angeli, G.; Deinhammer, H.; Faderl, J.; Angeli, J.; Papst, I. and Wolpers, M. Distribution
and Composition of Surface Oxides on High Strength IF-Steels after Recrystallization
Annealing. 44th MWSP Conference Proceedings, vol. XL, pp. 825-832, 2002
49. Jordan, C. E. and Marder, A. R. Fe-Zn Phase Formation in Interstitial-Free Steels Hot-
Dip Galvanized at 450 0C. Journal of Materials Science 32 (1997), pp. 5593-5602, 1997
108
50. Feliu Jr., S. and Pérez-Revenga, M. L. Correlation between the Surface Chemistry of
Annealed IF Steels and the Growth of a Galvanneal Coating. Acta Materialia, March 2005
51. Guttman, M. Diffusive Phase Transformations in Hot Dip Galvanizing. Materials Science
Forum vols. 155-156 (1994), pp. 527-548, 1994
52. Fujibayashi, N.; Kyono, K.; Kato, C. and Morito, N. Influence of Carbide-Forming-
Elements in Substrate Steels on Galvannealing Behavior. Proceedings of the 4th
International Conference on Zinc and Zinc Alloy Coated Steel Sheet (GALVATECH’98),
Chiba, Japan, The Iron and Steel Institute of Japan, 1998
53. Cheng. C. Galvannealing Behavior and Coating Performance of Ti-Nb and Nb-Ti-IF
Steels. 42nd MWSP Conference Proceedings, ISS, vol. XXXVIII, pp. 255-263, 2000
54. Shindo, Y. Influence of Precipitates in IF Steel on Galvanneal Coatings Structure. 42nd
MWSP Conference Proceedings, ISS, vol. XXXVIII, pp. 265-276, 2000
55. Miner, R. E. and Butler, J. F. and Dunbar, F. C. Comparison of Product Characteristics
of Electrogalvanized and Galvannealed for Exposed Automotive Applications. Proceedings
of the 3rd International Conference on Zinc and Zinc Alloy Coated Steel Sheet
(GALVATECH’95), pp. 407-412, 1995
56. Jordan, C. E. and Marder, A. R. Inhibition Layer Breakdown and Outburst Fe-Zn Alloy.
Zinc-Based Steel Coating Systems: Production and Performance, pp. 115-126, 1998
57. DeArdo, A. J. Physical Metallurgy of Interstitial-Free Steels: Precipitates and Solutes. IF
Steels 2000 Proceedings, ISS-AIME, pp. 125-136, 2000
58. Kino. N.; Yamada, M.; Tokunaga, Y. and Tsuchiya, H. Production of Nb-Ti-added Ultra-
Low-Carbon Steel for Galvannealed Application. Metallurgy of Vacuum Degassed Steel
Products, pp. 197-213, 1990
109
59. Tither, G.; Garcia, C. I.; Hua, M. and DeArdo, A. J. Precipitation Behavior and Solute
Effects in Interstitial-Free Steels. Physical Metallurgy of IF Steels, ISIJ, pp. 292-322, 1994
60. Puente, J. M.; Alonso, F.; Andrés, L. and Prado, M. Influence of an Adequate Surface
Conditioning on the Final Characteristics of GI for Exposed Panels Use on Automotive
Sector. Proceedings of the 6th International Conference on Zinc and Zinc Alloy Coated
Sheet Steels (GALVATECH’04), pp. 457-463, Chicago, 2004
61. Chen, T. R.; Hua, M.; Garcia, C. I. And DeArdo, A. J. Role of Niobium on the
Segregation Behavior of Phosphorus and Manganese in High Strength IF Sheet Steels for
Automotive Use. Developments in Sheet Products for Automotive Applications, Materials
Science & Technology 2005, pp. 163-171, 2005
62. Grabke, H. J.; Leroy, V. and Viefhaus, H. Segregation on the Surface of Steels in Heat
Treatment and Oxidation. ISIJ International, vol. 35 (1995), no. 2, pp. 95-113, 1995
63. Leprêtre, Y.; Mataigne, J. M.; Guttmann, M. and Philibert, J. Reactive Interdiffusion in
the Fe-Al-Zn System: Reaction Mechanisms during Hot-Dip Galvanizing. Zinc-Based Steel
Coating Systems: Production and Performance, pp. 95-106, 1998
64. Misra, R. D. K. Temperature-Time Dependence of Grain Boundary Segregation of
Phosphorus in Interstitial-Free Steels. Journal of Materials Science Letters 21, pp. 1275-
1279, 2002
65. Tang, N. Y. Thermodynamics and Kinetics of Alloy Formation in Galvanized Coatings.
Zinc-Based Steel Coating Systems: Production and Performance, pp. 3-11, 1998
66. Giorgi, M. L. and Guillot, J. B. Modeling of the Kinetics of Galvanizing Reactions.
Proceedings of the 6th International Conference on Zinc and Zinc Alloy Coated Steel Sheet
(GALVATECH’04), pp. 703-711, Chicago, 2004
110
67. Leroy, V. Surface Reaction of Steels in Batch Annealing: Relation with Residual Carbon
Contamination and Partial Selective Oxidation. 1st Internal Conference HICON-H2. Linz,
1992.
68. Wagner, C. Elektrochemie, 63 (1959), 772.
69. Birks, N. and Meier, G. H. Introduction to High Temperature Oxidation of Metals.
Edward Arnold Publishers. London, 1983.
70. Kaushik, P; Kruse, D.; Ozgu, M. Assessment of Castability Issues in Interstitial Free
Steels. La Revue de Metallurgie - CIT, Février 2008, pp. 92 – 101.
71. Bordignon, L.; Vanden Eynde, X. Zinc Wetting During Hot Dip Galvanizing. La Revue
de Métallurgie-CIT, Juin 2007, pp. 300 – 307.
72. Mizui, N. Precipitation Control and Related Mechanical Property in Ultra-Low Carbon
Sheet Steels. International Symposium on Modern LC and ULC Sheet Steels for Cold
Forming: Processing and Properties. Vol. 1, pp. 169-178. Aachen, 1998.